From karl at iwl.com Mon Jun 1 11:22:51 2026 From: karl at iwl.com (Karl Auerbach) Date: Mon, 1 Jun 2026 11:22:51 -0700 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club Message-ID: This is a trivial, and quite irreverent, bit of Internet history.... I, like several others, were members of the UCLA computer club during the late 1960's.? The club's office was in Boelter Hall - not far from the room that held IMP #1 (and the Sigma computer - along with its "Sigma EXecutive" documentation, aka "SEX Manuals".? My project's computer, an IBM 7094 - with a true memory leak [the core memory was oil cooled, and that oil leaked] - was in the next room over and we could hear the squeals from the AM radio caused by the RF noise from the IMP and the Sigma.) Anyway, folks in the Computer Club - especially Mark Kampe - kinda like to pull pranks.? For instance, we would drop things from the top of Boelter Hall (9 floors up) to see what would happen.? The landing zone was the collection of crunched and bent automobiles resulting from the early crash-tests of my group, the Institute of Traffic and Traffic Engineering.? We dropped everything from frozen superballs to a lead container used to hold/transport radioactive materials [it was empty].? There was also a feisty ice-cream vending machine in the hallway that once-too-often failed to deliver the paid-for frozen treat - so someone in the club unplugged the machine for a few hours, everything inside melted, and then plugged it back in, re-frezzing the leaking drippy mess.? That was not nice, but it was - here's a terrible pun - that vending machine received its just desserts. Anyway, back to Internet History... IMP #1 had the rough appearance of an armor plated refrigerator, with lifting lugs on the top.? The machine was "ruggedized". That word, ruggedized, was like honey to ants - it seriously caught our attention.? So we (I think Mark K. in particular) asked "Is it rugged enough to survive a drop from the top of Boelter Hall?". So our imaginations lit up with images of us grabbing IMP #1, hauling it up to the roof and dropping it into the crashed cars nine floors below. Obviously, prudence and sanity - and perhaps even some, probably reluctant, respect for law - prevailed.? We never did get beyond the "what if we did this" stage. (But a couple of years later some of us migrated from UCLA to SDC in Santa Monica.? At SDC we had an extremely awful HP 2000 minicomputer that, if I remember properly, did actually suffer such a fate as it was dropped it from the roof of the Q7A building - three stories tall - onto the parking lot - a fate that all of us applauded. [It was truly a terrible machine with an even worse operating system.]? Some of use, years later, moved onto the Interop show nets were we sometime had to practice the delicate art of percussive maintenance.) ? ? ? ? --karl-- From jack at 3kitty.org Mon Jun 1 11:41:23 2026 From: jack at 3kitty.org (Jack Haverty) Date: Mon, 1 Jun 2026 11:41:23 -0700 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: References: Message-ID: Another IMP story... The "Pluribus" computer was a BBN creation that included highly redundant hardware, with multiple everythings.? ?The idea was that no matter what failed, the system would keep running while repairs were made.? Of course a Pluribus could be used as an IMP, which was popular in some government installations.? One day at BBN, I heard this story.? Wasn't there myself, but I can believe it. A Pluribus IMP was being decommissioned at some government site. ?They happened to have a variety of military stuff around.? ?So someone decided to see if the Pluribus IMP was as reliable as it was touted to be. The IMP was set up, still running.? ?Someone got a rifle (M-16?) and started shooting at the IMP.? Really.? Sadly I don't recall the number, but the IMP survived an amazing number of direct hits at point blank range, and still kept passing traffic. Where did those folks in the Computer Club go after leaving UCLA...?? Any of them in ROTC? /Jack On 6/1/26 11:22, Karl Auerbach via Internet-history wrote: > This is a trivial, and quite irreverent, bit of Internet history.... > > I, like several others, were members of the UCLA computer club during > the late 1960's.? The club's office was in Boelter Hall - not far from > the room that held IMP #1 (and the Sigma computer - along with its > "Sigma EXecutive" documentation, aka "SEX Manuals".? My project's > computer, an IBM 7094 - with a true memory leak [the core memory was > oil cooled, and that oil leaked] - was in the next room over and we > could hear the squeals from the AM radio caused by the RF noise from > the IMP and the Sigma.) > > Anyway, folks in the Computer Club - especially Mark Kampe - kinda > like to pull pranks.? For instance, we would drop things from the top > of Boelter Hall (9 floors up) to see what would happen.? The landing > zone was the collection of crunched and bent automobiles resulting > from the early crash-tests of my group, the Institute of Traffic and > Traffic Engineering.? We dropped everything from frozen superballs to > a lead container used to hold/transport radioactive materials [it was > empty].? There was also a feisty ice-cream vending machine in the > hallway that once-too-often failed to deliver the paid-for frozen > treat - so someone in the club unplugged the machine for a few hours, > everything inside melted, and then plugged it back in, re-frezzing the > leaking drippy mess.? That was not nice, but it was - here's a > terrible pun - that vending machine received its just desserts. > > Anyway, back to Internet History... > > IMP #1 had the rough appearance of an armor plated refrigerator, with > lifting lugs on the top.? The machine was "ruggedized". > > That word, ruggedized, was like honey to ants - it seriously caught > our attention.? So we (I think Mark K. in particular) asked "Is it > rugged enough to survive a drop from the top of Boelter Hall?". > > So our imaginations lit up with images of us grabbing IMP #1, hauling > it up to the roof and dropping it into the crashed cars nine floors > below. > > Obviously, prudence and sanity - and perhaps even some, probably > reluctant, respect for law - prevailed.? We never did get beyond the > "what if we did this" stage. > > (But a couple of years later some of us migrated from UCLA to SDC in > Santa Monica.? At SDC we had an extremely awful HP 2000 minicomputer > that, if I remember properly, did actually suffer such a fate as it > was dropped it from the roof of the Q7A building - three stories tall > - onto the parking lot - a fate that all of us applauded. [It was > truly a terrible machine with an even worse operating system.]? Some > of use, years later, moved onto the Interop show nets were we sometime > had to practice the delicate art of percussive maintenance.) > > ? ? ? ? --karl-- > -------------- next part -------------- A non-text attachment was scrubbed... Name: OpenPGP_signature.asc Type: application/pgp-signature Size: 665 bytes Desc: OpenPGP digital signature URL: From steve at shinkuro.com Mon Jun 1 11:52:01 2026 From: steve at shinkuro.com (Steve Crocker) Date: Mon, 1 Jun 2026 14:52:01 -0400 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: References: Message-ID: I was vice president and the president of the UCLA computer club 1961-1962. Steve Sent from my iPhone > On Jun 1, 2026, at 2:41?PM, Jack Haverty via Internet-history wrote: > > ?Another IMP story... > > The "Pluribus" computer was a BBN creation that included highly redundant hardware, with multiple everythings. The idea was that no matter what failed, the system would keep running while repairs were made. Of course a Pluribus could be used as an IMP, which was popular in some government installations. One day at BBN, I heard this story. Wasn't there myself, but I can believe it. > > A Pluribus IMP was being decommissioned at some government site. They happened to have a variety of military stuff around. So someone decided to see if the Pluribus IMP was as reliable as it was touted to be. > > The IMP was set up, still running. Someone got a rifle (M-16?) and started shooting at the IMP. Really. Sadly I don't recall the number, but the IMP survived an amazing number of direct hits at point blank range, and still kept passing traffic. > > Where did those folks in the Computer Club go after leaving UCLA...? Any of them in ROTC? > > /Jack > >> On 6/1/26 11:22, Karl Auerbach via Internet-history wrote: >> This is a trivial, and quite irreverent, bit of Internet history.... >> >> I, like several others, were members of the UCLA computer club during the late 1960's. The club's office was in Boelter Hall - not far from the room that held IMP #1 (and the Sigma computer - along with its "Sigma EXecutive" documentation, aka "SEX Manuals". My project's computer, an IBM 7094 - with a true memory leak [the core memory was oil cooled, and that oil leaked] - was in the next room over and we could hear the squeals from the AM radio caused by the RF noise from the IMP and the Sigma.) >> >> Anyway, folks in the Computer Club - especially Mark Kampe - kinda like to pull pranks. For instance, we would drop things from the top of Boelter Hall (9 floors up) to see what would happen. The landing zone was the collection of crunched and bent automobiles resulting from the early crash-tests of my group, the Institute of Traffic and Traffic Engineering. We dropped everything from frozen superballs to a lead container used to hold/transport radioactive materials [it was empty]. There was also a feisty ice-cream vending machine in the hallway that once-too-often failed to deliver the paid-for frozen treat - so someone in the club unplugged the machine for a few hours, everything inside melted, and then plugged it back in, re-frezzing the leaking drippy mess. That was not nice, but it was - here's a terrible pun - that vending machine received its just desserts. >> >> Anyway, back to Internet History... >> >> IMP #1 had the rough appearance of an armor plated refrigerator, with lifting lugs on the top. The machine was "ruggedized". >> >> That word, ruggedized, was like honey to ants - it seriously caught our attention. So we (I think Mark K. in particular) asked "Is it rugged enough to survive a drop from the top of Boelter Hall?". >> >> So our imaginations lit up with images of us grabbing IMP #1, hauling it up to the roof and dropping it into the crashed cars nine floors below. >> >> Obviously, prudence and sanity - and perhaps even some, probably reluctant, respect for law - prevailed. We never did get beyond the "what if we did this" stage. >> >> (But a couple of years later some of us migrated from UCLA to SDC in Santa Monica. At SDC we had an extremely awful HP 2000 minicomputer that, if I remember properly, did actually suffer such a fate as it was dropped it from the roof of the Q7A building - three stories tall - onto the parking lot - a fate that all of us applauded. [It was truly a terrible machine with an even worse operating system.] Some of use, years later, moved onto the Interop show nets were we sometime had to practice the delicate art of percussive maintenance.) >> >> --karl-- >> > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From karl at iwl.com Mon Jun 1 12:18:10 2026 From: karl at iwl.com (Karl Auerbach) Date: Mon, 1 Jun 2026 12:18:10 -0700 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: References: Message-ID: On 6/1/26 11:41 AM, Jack Haverty via Internet-history wrote: > > Where did those folks in the Computer Club go after leaving UCLA...?? > Any of them in ROTC? Well, I suspect that some of us ended up getting drafted into the military.? (I almost was.) Many of us ended up at Rand or SDC - only a few miles west of UCLA. Regarding the story of the Plurebus ... at SDC I was part of the research group for operating system and network security.? We deeply explored capability architectures.? Such architectures are good not only for security but also for isolating and limiting the impact of both hardware failures and software errors.? One of our Plessy machines - a machine intended for use in a telco switching environment - was architected to have no more than ten minutes of downtime over a forty year span - including down time due to software errors.? (I got to design the hardware for a capability machine to run Peter Neumann's PSOS? - Provably Secure Operating System - we had to make significant architectural changes PSOS to create a machine that actually would work, most particularly changing from capabilities with infinite lifetimes to capabilities that were finite and could be garbage collected - Peter and I carried on that debate for decades afterwords.? I think he may have conceded in CHERI architecture.) For a while I was working at RSRE - Royal Signals and Radar Establishment - in Great Malvern, UK.??(Sometimes when I had to do something in the Slough or Maidenhead facilities I got to work a little bit with Donald Davies.) RSRE was an RAF facility - surprisingly, with a copy of Picasso's "Guernica" on the wall of the officer's mess hall.? I, being utterly naive about military ranks and procedures, did not realize that the typical rank of the people with whom I was working - all had masters or phd's - was colonel.? And I did not realize that colonel was a fairly serious rank that could bend rules. So one afternoon - after our daily game of croquette after lunch - one of the RAF colonels said "You say that this machine is robust and will continue to run.? Hmmm, let me think upon this." He then pulled a hand grenade out of his kit and mimed pulling the pin and tossing it into the machine.? We conceded that that kind of thing was most likely beyond the survival abilities of the machine. ? ? ? ? --karl-- From craig at tereschau.net Mon Jun 1 12:40:06 2026 From: craig at tereschau.net (Craig Partridge) Date: Mon, 1 Jun 2026 13:40:06 -0600 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: References: Message-ID: By the time I joined BBN, the story was that the person shooting the Pluribus had been forced to write programs for it. It had a BBN-wide reputation for not being programmer-friendly (which is saying something, as BBN periodically generated computing platforms which were painful to program -- such as the C70 with 10-bit bytes). Craig On Mon, Jun 1, 2026 at 12:41?PM Jack Haverty via Internet-history < internet-history at elists.isoc.org> wrote: > Another IMP story... > > The "Pluribus" computer was a BBN creation that included highly > redundant hardware, with multiple everythings. The idea was that no > matter what failed, the system would keep running while repairs were > made. Of course a Pluribus could be used as an IMP, which was popular > in some government installations. One day at BBN, I heard this story. > Wasn't there myself, but I can believe it. > > A Pluribus IMP was being decommissioned at some government site. They > happened to have a variety of military stuff around. So someone > decided to see if the Pluribus IMP was as reliable as it was touted to be. > > The IMP was set up, still running. Someone got a rifle (M-16?) and > started shooting at the IMP. Really. Sadly I don't recall the number, > but the IMP survived an amazing number of direct hits at point blank > range, and still kept passing traffic. > > Where did those folks in the Computer Club go after leaving UCLA...? > Any of them in ROTC? > > /Jack > > On 6/1/26 11:22, Karl Auerbach via Internet-history wrote: > > This is a trivial, and quite irreverent, bit of Internet history.... > > > > I, like several others, were members of the UCLA computer club during > > the late 1960's. The club's office was in Boelter Hall - not far from > > the room that held IMP #1 (and the Sigma computer - along with its > > "Sigma EXecutive" documentation, aka "SEX Manuals". My project's > > computer, an IBM 7094 - with a true memory leak [the core memory was > > oil cooled, and that oil leaked] - was in the next room over and we > > could hear the squeals from the AM radio caused by the RF noise from > > the IMP and the Sigma.) > > > > Anyway, folks in the Computer Club - especially Mark Kampe - kinda > > like to pull pranks. For instance, we would drop things from the top > > of Boelter Hall (9 floors up) to see what would happen. The landing > > zone was the collection of crunched and bent automobiles resulting > > from the early crash-tests of my group, the Institute of Traffic and > > Traffic Engineering. We dropped everything from frozen superballs to > > a lead container used to hold/transport radioactive materials [it was > > empty]. There was also a feisty ice-cream vending machine in the > > hallway that once-too-often failed to deliver the paid-for frozen > > treat - so someone in the club unplugged the machine for a few hours, > > everything inside melted, and then plugged it back in, re-frezzing the > > leaking drippy mess. That was not nice, but it was - here's a > > terrible pun - that vending machine received its just desserts. > > > > Anyway, back to Internet History... > > > > IMP #1 had the rough appearance of an armor plated refrigerator, with > > lifting lugs on the top. The machine was "ruggedized". > > > > That word, ruggedized, was like honey to ants - it seriously caught > > our attention. So we (I think Mark K. in particular) asked "Is it > > rugged enough to survive a drop from the top of Boelter Hall?". > > > > So our imaginations lit up with images of us grabbing IMP #1, hauling > > it up to the roof and dropping it into the crashed cars nine floors > > below. > > > > Obviously, prudence and sanity - and perhaps even some, probably > > reluctant, respect for law - prevailed. We never did get beyond the > > "what if we did this" stage. > > > > (But a couple of years later some of us migrated from UCLA to SDC in > > Santa Monica. At SDC we had an extremely awful HP 2000 minicomputer > > that, if I remember properly, did actually suffer such a fate as it > > was dropped it from the roof of the Q7A building - three stories tall > > - onto the parking lot - a fate that all of us applauded. [It was > > truly a terrible machine with an even worse operating system.] Some > > of use, years later, moved onto the Interop show nets were we sometime > > had to practice the delicate art of percussive maintenance.) > > > > --karl-- > > > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- ***** Craig Partridge's email account for professional society activities and mailing lists. From agmalis at gmail.com Mon Jun 1 14:21:15 2026 From: agmalis at gmail.com (Andrew G. Malis) Date: Mon, 1 Jun 2026 17:21:15 -0400 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: References: Message-ID: Craig, I did my share of Pluribus coding and debugging. The assembly code was really no worse than any other once you got the hang of it, it was actually somewhat similar to PDP-11 assembler. It was a multiprocessor machine, and if any of the processors crashed, the remaining processors would notice, send a fresh code reload to the failed processor, and restart it. As a result, it was a pain to try to debug issues in the field, although crashes would produce dumps for later analysis. The easiest way to debug the code in the lab was to shut down all but one processor. We were able to shut down (and keep down) individual processors by using the front panel switches to write "FADE" into a particular location. We could later reboot the FADEd processor(s) manually. Regarding physically stopping a Pluribus by bullets or hand grenade, I certainly believe a hand grenade would damage all of the processors enough to bring them all down. One well-placed bullet in each processor would probably do it as well, but you would have to do it carefully, not just fire willy-nilly at the collection of processors, you could easily miss at least one of the processors that way. Cheers, Andy On Mon, Jun 1, 2026 at 3:40?PM Craig Partridge via Internet-history < internet-history at elists.isoc.org> wrote: > By the time I joined BBN, the story was that the person shooting the > Pluribus had been forced to write programs for it. It had a BBN-wide > reputation for not being programmer-friendly (which is saying something, as > BBN periodically generated computing platforms which were painful to > program -- such as the C70 with 10-bit bytes). > > Craig > > On Mon, Jun 1, 2026 at 12:41?PM Jack Haverty via Internet-history < > internet-history at elists.isoc.org> wrote: > > > Another IMP story... > > > > The "Pluribus" computer was a BBN creation that included highly > > redundant hardware, with multiple everythings. The idea was that no > > matter what failed, the system would keep running while repairs were > > made. Of course a Pluribus could be used as an IMP, which was popular > > in some government installations. One day at BBN, I heard this story. > > Wasn't there myself, but I can believe it. > > > > A Pluribus IMP was being decommissioned at some government site. They > > happened to have a variety of military stuff around. So someone > > decided to see if the Pluribus IMP was as reliable as it was touted to > be. > > > > The IMP was set up, still running. Someone got a rifle (M-16?) and > > started shooting at the IMP. Really. Sadly I don't recall the number, > > but the IMP survived an amazing number of direct hits at point blank > > range, and still kept passing traffic. > > > > Where did those folks in the Computer Club go after leaving UCLA...? > > Any of them in ROTC? > > > > /Jack > > > > On 6/1/26 11:22, Karl Auerbach via Internet-history wrote: > > > This is a trivial, and quite irreverent, bit of Internet history.... > > > > > > I, like several others, were members of the UCLA computer club during > > > the late 1960's. The club's office was in Boelter Hall - not far from > > > the room that held IMP #1 (and the Sigma computer - along with its > > > "Sigma EXecutive" documentation, aka "SEX Manuals". My project's > > > computer, an IBM 7094 - with a true memory leak [the core memory was > > > oil cooled, and that oil leaked] - was in the next room over and we > > > could hear the squeals from the AM radio caused by the RF noise from > > > the IMP and the Sigma.) > > > > > > Anyway, folks in the Computer Club - especially Mark Kampe - kinda > > > like to pull pranks. For instance, we would drop things from the top > > > of Boelter Hall (9 floors up) to see what would happen. The landing > > > zone was the collection of crunched and bent automobiles resulting > > > from the early crash-tests of my group, the Institute of Traffic and > > > Traffic Engineering. We dropped everything from frozen superballs to > > > a lead container used to hold/transport radioactive materials [it was > > > empty]. There was also a feisty ice-cream vending machine in the > > > hallway that once-too-often failed to deliver the paid-for frozen > > > treat - so someone in the club unplugged the machine for a few hours, > > > everything inside melted, and then plugged it back in, re-frezzing the > > > leaking drippy mess. That was not nice, but it was - here's a > > > terrible pun - that vending machine received its just desserts. > > > > > > Anyway, back to Internet History... > > > > > > IMP #1 had the rough appearance of an armor plated refrigerator, with > > > lifting lugs on the top. The machine was "ruggedized". > > > > > > That word, ruggedized, was like honey to ants - it seriously caught > > > our attention. So we (I think Mark K. in particular) asked "Is it > > > rugged enough to survive a drop from the top of Boelter Hall?". > > > > > > So our imaginations lit up with images of us grabbing IMP #1, hauling > > > it up to the roof and dropping it into the crashed cars nine floors > > > below. > > > > > > Obviously, prudence and sanity - and perhaps even some, probably > > > reluctant, respect for law - prevailed. We never did get beyond the > > > "what if we did this" stage. > > > > > > (But a couple of years later some of us migrated from UCLA to SDC in > > > Santa Monica. At SDC we had an extremely awful HP 2000 minicomputer > > > that, if I remember properly, did actually suffer such a fate as it > > > was dropped it from the roof of the Q7A building - three stories tall > > > - onto the parking lot - a fate that all of us applauded. [It was > > > truly a terrible machine with an even worse operating system.] Some > > > of use, years later, moved onto the Interop show nets were we sometime > > > had to practice the delicate art of percussive maintenance.) > > > > > > --karl-- > > > > > > > -- > > Internet-history mailing list > > Internet-history at elists.isoc.org > > https://elists.isoc.org/mailman/listinfo/internet-history > > - > > Unsubscribe: > > > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > > > > -- > ***** > Craig Partridge's email account for professional society activities and > mailing lists. > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From vint at google.com Mon Jun 1 14:37:19 2026 From: vint at google.com (Vint Cerf) Date: Mon, 1 Jun 2026 17:37:19 -0400 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: References: Message-ID: While we are in anecdote mode, the Sigma-7, like other processors, could be used to toggle a flip-flop at various rates to produce sound. This was usually accomplished by putting the machine into a loop with a certain repetition rate. One of the senior graduate students was into computer music. One day, I brought a friend to see the Sigma-7 and the IMP. When we arrived, the sr. grad. had a set of tuning forks and was adjusting the loop lengths to match the sounds of the tuning forks. As we approached, I cautioned my guest to be quiet. "He's tuning the computer, I said...." v On Mon, Jun 1, 2026 at 5:21?PM Andrew G. Malis via Internet-history < internet-history at elists.isoc.org> wrote: > Craig, > > I did my share of Pluribus coding and debugging. The assembly code was > really no worse than any other once you got the hang of it, it was actually > somewhat similar to PDP-11 assembler. It was a multiprocessor machine, and > if any of the processors crashed, the remaining processors would notice, > send a fresh code reload to the failed processor, and restart it. As a > result, it was a pain to try to debug issues in the field, although crashes > would produce dumps for later analysis. The easiest way to debug the code > in the lab was to shut down all but one processor. We were able to shut > down (and keep down) individual processors by using the front panel > switches to write "FADE" into a particular location. We could later reboot > the FADEd processor(s) manually. > > Regarding physically stopping a Pluribus by bullets or hand grenade, I > certainly believe a hand grenade would damage all of the processors enough > to bring them all down. One well-placed bullet in each processor would > probably do it as well, but you would have to do it carefully, not just > fire willy-nilly at the collection of processors, you could easily miss at > least one of the processors that way. > > Cheers, > Andy > > > > On Mon, Jun 1, 2026 at 3:40?PM Craig Partridge via Internet-history < > internet-history at elists.isoc.org> wrote: > > > By the time I joined BBN, the story was that the person shooting the > > Pluribus had been forced to write programs for it. It had a BBN-wide > > reputation for not being programmer-friendly (which is saying something, > as > > BBN periodically generated computing platforms which were painful to > > program -- such as the C70 with 10-bit bytes). > > > > Craig > > > > On Mon, Jun 1, 2026 at 12:41?PM Jack Haverty via Internet-history < > > internet-history at elists.isoc.org> wrote: > > > > > Another IMP story... > > > > > > The "Pluribus" computer was a BBN creation that included highly > > > redundant hardware, with multiple everythings. The idea was that no > > > matter what failed, the system would keep running while repairs were > > > made. Of course a Pluribus could be used as an IMP, which was popular > > > in some government installations. One day at BBN, I heard this story. > > > Wasn't there myself, but I can believe it. > > > > > > A Pluribus IMP was being decommissioned at some government site. They > > > happened to have a variety of military stuff around. So someone > > > decided to see if the Pluribus IMP was as reliable as it was touted to > > be. > > > > > > The IMP was set up, still running. Someone got a rifle (M-16?) and > > > started shooting at the IMP. Really. Sadly I don't recall the number, > > > but the IMP survived an amazing number of direct hits at point blank > > > range, and still kept passing traffic. > > > > > > Where did those folks in the Computer Club go after leaving UCLA...? > > > Any of them in ROTC? > > > > > > /Jack > > > > > > On 6/1/26 11:22, Karl Auerbach via Internet-history wrote: > > > > This is a trivial, and quite irreverent, bit of Internet history.... > > > > > > > > I, like several others, were members of the UCLA computer club during > > > > the late 1960's. The club's office was in Boelter Hall - not far > from > > > > the room that held IMP #1 (and the Sigma computer - along with its > > > > "Sigma EXecutive" documentation, aka "SEX Manuals". My project's > > > > computer, an IBM 7094 - with a true memory leak [the core memory was > > > > oil cooled, and that oil leaked] - was in the next room over and we > > > > could hear the squeals from the AM radio caused by the RF noise from > > > > the IMP and the Sigma.) > > > > > > > > Anyway, folks in the Computer Club - especially Mark Kampe - kinda > > > > like to pull pranks. For instance, we would drop things from the top > > > > of Boelter Hall (9 floors up) to see what would happen. The landing > > > > zone was the collection of crunched and bent automobiles resulting > > > > from the early crash-tests of my group, the Institute of Traffic and > > > > Traffic Engineering. We dropped everything from frozen superballs to > > > > a lead container used to hold/transport radioactive materials [it was > > > > empty]. There was also a feisty ice-cream vending machine in the > > > > hallway that once-too-often failed to deliver the paid-for frozen > > > > treat - so someone in the club unplugged the machine for a few hours, > > > > everything inside melted, and then plugged it back in, re-frezzing > the > > > > leaking drippy mess. That was not nice, but it was - here's a > > > > terrible pun - that vending machine received its just desserts. > > > > > > > > Anyway, back to Internet History... > > > > > > > > IMP #1 had the rough appearance of an armor plated refrigerator, with > > > > lifting lugs on the top. The machine was "ruggedized". > > > > > > > > That word, ruggedized, was like honey to ants - it seriously caught > > > > our attention. So we (I think Mark K. in particular) asked "Is it > > > > rugged enough to survive a drop from the top of Boelter Hall?". > > > > > > > > So our imaginations lit up with images of us grabbing IMP #1, hauling > > > > it up to the roof and dropping it into the crashed cars nine floors > > > > below. > > > > > > > > Obviously, prudence and sanity - and perhaps even some, probably > > > > reluctant, respect for law - prevailed. We never did get beyond the > > > > "what if we did this" stage. > > > > > > > > (But a couple of years later some of us migrated from UCLA to SDC in > > > > Santa Monica. At SDC we had an extremely awful HP 2000 minicomputer > > > > that, if I remember properly, did actually suffer such a fate as it > > > > was dropped it from the roof of the Q7A building - three stories tall > > > > - onto the parking lot - a fate that all of us applauded. [It was > > > > truly a terrible machine with an even worse operating system.] Some > > > > of use, years later, moved onto the Interop show nets were we > sometime > > > > had to practice the delicate art of percussive maintenance.) > > > > > > > > --karl-- > > > > > > > > > > -- > > > Internet-history mailing list > > > Internet-history at elists.isoc.org > > > https://elists.isoc.org/mailman/listinfo/internet-history > > > - > > > Unsubscribe: > > > > > > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > > > > > > > > -- > > ***** > > Craig Partridge's email account for professional society activities and > > mailing lists. > > -- > > Internet-history mailing list > > Internet-history at elists.isoc.org > > https://elists.isoc.org/mailman/listinfo/internet-history > > - > > Unsubscribe: > > > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- After July 1, Please send any postal/overnight deliveries to: Vint Cerf 1435 Woodhurst Blvd McLean, VA 22102 Until July 1, send postal/overnight deliveries to: Google, LLC 1900 Reston Metro Plaza, 16th Floor Reston, VA 20190 +1 650-224-2788 From jeanjour at comcast.net Mon Jun 1 16:28:45 2026 From: jeanjour at comcast.net (John Day) Date: Mon, 1 Jun 2026 19:28:45 -0400 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: References: Message-ID: <0CCE4623-49CE-4BCF-AEC6-6F09178FA36D@comcast.net> We once convinced a poor undergrad that every night, the operator had to empty the bit-bucket from all of the left and right shifts during the day. To convince the poor student, they were shown the card chad from a keypunch no one used. > On Jun 1, 2026, at 17:37, Vint Cerf via Internet-history wrote: > > While we are in anecdote mode, the Sigma-7, like other processors, could be > used to toggle a flip-flop at various rates to produce sound. This was > usually accomplished by putting the machine into a loop with a certain > repetition rate. One of the senior graduate students was into computer > music. One day, I brought a friend to see the Sigma-7 and the IMP. When we > arrived, the sr. grad. had a set of tuning forks and was adjusting the loop > lengths to match the sounds of the tuning forks. As we approached, I > cautioned my guest to be quiet. "He's tuning the computer, I said...." > > v > > > On Mon, Jun 1, 2026 at 5:21?PM Andrew G. Malis via Internet-history < > internet-history at elists.isoc.org> wrote: > >> Craig, >> >> I did my share of Pluribus coding and debugging. The assembly code was >> really no worse than any other once you got the hang of it, it was actually >> somewhat similar to PDP-11 assembler. It was a multiprocessor machine, and >> if any of the processors crashed, the remaining processors would notice, >> send a fresh code reload to the failed processor, and restart it. As a >> result, it was a pain to try to debug issues in the field, although crashes >> would produce dumps for later analysis. The easiest way to debug the code >> in the lab was to shut down all but one processor. We were able to shut >> down (and keep down) individual processors by using the front panel >> switches to write "FADE" into a particular location. We could later reboot >> the FADEd processor(s) manually. >> >> Regarding physically stopping a Pluribus by bullets or hand grenade, I >> certainly believe a hand grenade would damage all of the processors enough >> to bring them all down. One well-placed bullet in each processor would >> probably do it as well, but you would have to do it carefully, not just >> fire willy-nilly at the collection of processors, you could easily miss at >> least one of the processors that way. >> >> Cheers, >> Andy >> >> >> >> On Mon, Jun 1, 2026 at 3:40?PM Craig Partridge via Internet-history < >> internet-history at elists.isoc.org> wrote: >> >>> By the time I joined BBN, the story was that the person shooting the >>> Pluribus had been forced to write programs for it. It had a BBN-wide >>> reputation for not being programmer-friendly (which is saying something, >> as >>> BBN periodically generated computing platforms which were painful to >>> program -- such as the C70 with 10-bit bytes). >>> >>> Craig >>> >>> On Mon, Jun 1, 2026 at 12:41?PM Jack Haverty via Internet-history < >>> internet-history at elists.isoc.org> wrote: >>> >>>> Another IMP story... >>>> >>>> The "Pluribus" computer was a BBN creation that included highly >>>> redundant hardware, with multiple everythings. The idea was that no >>>> matter what failed, the system would keep running while repairs were >>>> made. Of course a Pluribus could be used as an IMP, which was popular >>>> in some government installations. One day at BBN, I heard this story. >>>> Wasn't there myself, but I can believe it. >>>> >>>> A Pluribus IMP was being decommissioned at some government site. They >>>> happened to have a variety of military stuff around. So someone >>>> decided to see if the Pluribus IMP was as reliable as it was touted to >>> be. >>>> >>>> The IMP was set up, still running. Someone got a rifle (M-16?) and >>>> started shooting at the IMP. Really. Sadly I don't recall the number, >>>> but the IMP survived an amazing number of direct hits at point blank >>>> range, and still kept passing traffic. >>>> >>>> Where did those folks in the Computer Club go after leaving UCLA...? >>>> Any of them in ROTC? >>>> >>>> /Jack >>>> >>>> On 6/1/26 11:22, Karl Auerbach via Internet-history wrote: >>>>> This is a trivial, and quite irreverent, bit of Internet history.... >>>>> >>>>> I, like several others, were members of the UCLA computer club during >>>>> the late 1960's. The club's office was in Boelter Hall - not far >> from >>>>> the room that held IMP #1 (and the Sigma computer - along with its >>>>> "Sigma EXecutive" documentation, aka "SEX Manuals". My project's >>>>> computer, an IBM 7094 - with a true memory leak [the core memory was >>>>> oil cooled, and that oil leaked] - was in the next room over and we >>>>> could hear the squeals from the AM radio caused by the RF noise from >>>>> the IMP and the Sigma.) >>>>> >>>>> Anyway, folks in the Computer Club - especially Mark Kampe - kinda >>>>> like to pull pranks. For instance, we would drop things from the top >>>>> of Boelter Hall (9 floors up) to see what would happen. The landing >>>>> zone was the collection of crunched and bent automobiles resulting >>>>> from the early crash-tests of my group, the Institute of Traffic and >>>>> Traffic Engineering. We dropped everything from frozen superballs to >>>>> a lead container used to hold/transport radioactive materials [it was >>>>> empty]. There was also a feisty ice-cream vending machine in the >>>>> hallway that once-too-often failed to deliver the paid-for frozen >>>>> treat - so someone in the club unplugged the machine for a few hours, >>>>> everything inside melted, and then plugged it back in, re-frezzing >> the >>>>> leaking drippy mess. That was not nice, but it was - here's a >>>>> terrible pun - that vending machine received its just desserts. >>>>> >>>>> Anyway, back to Internet History... >>>>> >>>>> IMP #1 had the rough appearance of an armor plated refrigerator, with >>>>> lifting lugs on the top. The machine was "ruggedized". >>>>> >>>>> That word, ruggedized, was like honey to ants - it seriously caught >>>>> our attention. So we (I think Mark K. in particular) asked "Is it >>>>> rugged enough to survive a drop from the top of Boelter Hall?". >>>>> >>>>> So our imaginations lit up with images of us grabbing IMP #1, hauling >>>>> it up to the roof and dropping it into the crashed cars nine floors >>>>> below. >>>>> >>>>> Obviously, prudence and sanity - and perhaps even some, probably >>>>> reluctant, respect for law - prevailed. We never did get beyond the >>>>> "what if we did this" stage. >>>>> >>>>> (But a couple of years later some of us migrated from UCLA to SDC in >>>>> Santa Monica. At SDC we had an extremely awful HP 2000 minicomputer >>>>> that, if I remember properly, did actually suffer such a fate as it >>>>> was dropped it from the roof of the Q7A building - three stories tall >>>>> - onto the parking lot - a fate that all of us applauded. [It was >>>>> truly a terrible machine with an even worse operating system.] Some >>>>> of use, years later, moved onto the Interop show nets were we >> sometime >>>>> had to practice the delicate art of percussive maintenance.) >>>>> >>>>> --karl-- >>>>> >>>> >>>> -- >>>> Internet-history mailing list >>>> Internet-history at elists.isoc.org >>>> https://elists.isoc.org/mailman/listinfo/internet-history >>>> - >>>> Unsubscribe: >>>> >>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >>>> >>> >>> >>> -- >>> ***** >>> Craig Partridge's email account for professional society activities and >>> mailing lists. >>> -- >>> Internet-history mailing list >>> Internet-history at elists.isoc.org >>> https://elists.isoc.org/mailman/listinfo/internet-history >>> - >>> Unsubscribe: >>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >>> >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > > > -- > After July 1, Please send any postal/overnight deliveries to: > Vint Cerf > 1435 Woodhurst Blvd > McLean, VA 22102 > > Until July 1, send postal/overnight deliveries to: > Google, LLC > 1900 Reston Metro Plaza, 16th Floor > Reston, VA 20190 > +1 650-224-2788 > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From jnc at mercury.lcs.mit.edu Mon Jun 1 16:48:20 2026 From: jnc at mercury.lcs.mit.edu (Noel Chiappa) Date: Mon, 1 Jun 2026 19:48:20 -0400 (EDT) Subject: [ih] capacity v bandwidth Message-ID: <20260601234820.B548A18C074@mercury.lcs.mit.edu> > From: Jack Haverty > In the early Internet, the boxes interconnecting networks were called > "gateways". Today they're called "routers". But why the change...? > > So we started callig them "routers". Other companies (cisco, proteon, > ...) probably had similar experiences in their sales activities. If my memory isn't failing me (it well might be), I can take part of the blame. I do remember that I was pissed off because everyone and their brother (across the industry generally) called any box that did digital communications between two things a 'gateway'. E.g. a box that did email forwarding from BITNET to the Internet was called a 'gateway'. I.e. 'gateway' was useless as a technical term, because it covered an impossibly wide range of functionalities. (I am not sure if the p4200, the first Proteon router product, post-dated the 'gateway' -> 'router' change; I'd have to try and find an original manual. If it pre-dated, I may have taken Proteon experience into account too.) So I campaigned (I think it was me) in the IETF community to come up with a term limited to internetwork-level datagra packet switches, and 'router' was picked. I don't know if that change post-dated the creation of the IETF or not. I remember such large-scale questions (i.e. not within the purview of a WG, after Phill set up the WG structure) were often discussed on the main IETF mailing list, so if we still have the email archive from the start of that list, someone can dig into it. I remember that before the IETF existed, there was an email list (I think hosted at CNRI maybe, although CNRI didn't exist until 1986 - Jon's minutes of TCP/IP meetings stop at the end of 1980) where a lot of early TCP internetworking discussions ('TCP internetworking' since there must have been PUP internetworking discussions, too, inside Xerox) happened. Does anyone remember what it was called? Any technical history of the creation of TCP internetting would _really_ benefit from having access to that email archive (if it still exists somewhere; if not, maybe it would be possible to re-create it by picking through preserved emailboxes; or perhaps someone who printed out all their email still has those printouts). I feel sadly wary that a lot of our earliest history has been lost (since we didn't use physical memos, which many technical histories depend on for 'nuts and bolts' primary sources) - except for the copy stored in 'meat' CPUs (who will soon start dying off - historians take note, and act now, while you can). Noel From brian.e.carpenter at gmail.com Mon Jun 1 17:52:03 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Tue, 2 Jun 2026 12:52:03 +1200 Subject: [ih] capacity v bandwidth In-Reply-To: <20260601234820.B548A18C074@mercury.lcs.mit.edu> References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> Message-ID: <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> This is slightly complicated by the IGP/EGP terminology. That persisted even when "router" became commonplace. I think we can date it to between June 1988 and June 1989: RFC1058 (RIP, June 1988) uses "gateway" exclusively RFC1105 (the first version of BGP, June 1989) uses "router" exclusively, except in the name of the protocol! Wikipedia says that the p4200 came out in 1986. I couldn't find a manual, but its product name was in a May 1988 DoD report: "Proteon p4200 Gateway ... The p4200 gateway is a multiprotocol router, supporting (among other protocols) TCP/IP." [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf] Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) Regards/Ng? mihi Brian On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: > > From: Jack Haverty > > > In the early Internet, the boxes interconnecting networks were called > > "gateways". Today they're called "routers". But why the change...? > > > > So we started callig them "routers". Other companies (cisco, proteon, > > ...) probably had similar experiences in their sales activities. > > If my memory isn't failing me (it well might be), I can take part of the > blame. > > I do remember that I was pissed off because everyone and their brother > (across the industry generally) called any box that did digital > communications between two things a 'gateway'. E.g. a box that did email > forwarding from BITNET to the Internet was called a 'gateway'. I.e. 'gateway' > was useless as a technical term, because it covered an impossibly wide range > of functionalities. > > (I am not sure if the p4200, the first Proteon router product, > post-dated the 'gateway' -> 'router' change; I'd have to try and find an > original manual. If it pre-dated, I may have taken Proteon experience into > account too.) > > So I campaigned (I think it was me) in the IETF community to come up with a > term limited to internetwork-level datagra packet switches, and 'router' was > picked. > > > I don't know if that change post-dated the creation of the IETF or not. I > remember such large-scale questions (i.e. not within the purview of a WG, > after Phill set up the WG structure) were often discussed on the main IETF > mailing list, so if we still have the email archive from the start of that > list, someone can dig into it. > > I remember that before the IETF existed, there was an email list (I think > hosted at CNRI maybe, although CNRI didn't exist until 1986 - Jon's minutes > of TCP/IP meetings stop at the end of 1980) where a lot of early TCP > internetworking discussions ('TCP internetworking' since there must have been > PUP internetworking discussions, too, inside Xerox) happened. Does anyone > remember what it was called? > > Any technical history of the creation of TCP internetting would _really_ > benefit from having access to that email archive (if it still exists > somewhere; if not, maybe it would be possible to re-create it by picking > through preserved emailboxes; or perhaps someone who printed out all their > email still has those printouts). > > I feel sadly wary that a lot of our earliest history has been lost (since we > didn't use physical memos, which many technical histories depend on for 'nuts > and bolts' primary sources) - except for the copy stored in 'meat' CPUs (who > will soon start dying off - historians take note, and act now, while you can). > > Noel From vgcerf at gmail.com Mon Jun 1 18:32:02 2026 From: vgcerf at gmail.com (vinton cerf) Date: Mon, 1 Jun 2026 21:32:02 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> Message-ID: I had always thought that cisco introduced the term 'router" in 1984. v On Mon, Jun 1, 2026 at 8:52?PM Brian E Carpenter via Internet-history < internet-history at elists.isoc.org> wrote: > This is slightly complicated by the IGP/EGP terminology. That persisted > even when > "router" became commonplace. I think we can date it to between June 1988 > and June 1989: > > RFC1058 (RIP, June 1988) uses "gateway" exclusively > > RFC1105 (the first version of BGP, June 1989) uses "router" exclusively, > except in the name of the protocol! > > Wikipedia says that the p4200 came out in 1986. I couldn't find a manual, > but its product name was in a May 1988 DoD report: > > "Proteon p4200 Gateway > ... > The p4200 gateway is a multiprotocol router, supporting (among other > protocols) TCP/IP." > > [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf] > > Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) > > Regards/Ng? mihi > Brian > > On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: > > > From: Jack Haverty > > > > > In the early Internet, the boxes interconnecting networks were > called > > > "gateways". Today they're called "routers". But why the > change...? > > > > > > So we started callig them "routers". Other companies (cisco, > proteon, > > > ...) probably had similar experiences in their sales activities. > > > > If my memory isn't failing me (it well might be), I can take part of the > > blame. > > > > I do remember that I was pissed off because everyone and their brother > > (across the industry generally) called any box that did digital > > communications between two things a 'gateway'. E.g. a box that did email > > forwarding from BITNET to the Internet was called a 'gateway'. I.e. > 'gateway' > > was useless as a technical term, because it covered an impossibly wide > range > > of functionalities. > > > > (I am not sure if the p4200, the first Proteon router product, > > post-dated the 'gateway' -> 'router' change; I'd have to try and find an > > original manual. If it pre-dated, I may have taken Proteon experience > into > > account too.) > > > > So I campaigned (I think it was me) in the IETF community to come up > with a > > term limited to internetwork-level datagra packet switches, and 'router' > was > > picked. > > > > > > I don't know if that change post-dated the creation of the IETF or not. I > > remember such large-scale questions (i.e. not within the purview of a WG, > > after Phill set up the WG structure) were often discussed on the main > IETF > > mailing list, so if we still have the email archive from the start of > that > > list, someone can dig into it. > > > > I remember that before the IETF existed, there was an email list (I think > > hosted at CNRI maybe, although CNRI didn't exist until 1986 - Jon's > minutes > > of TCP/IP meetings stop at the end of 1980) where a lot of early TCP > > internetworking discussions ('TCP internetworking' since there must have > been > > PUP internetworking discussions, too, inside Xerox) happened. Does anyone > > remember what it was called? > > > > Any technical history of the creation of TCP internetting would _really_ > > benefit from having access to that email archive (if it still exists > > somewhere; if not, maybe it would be possible to re-create it by picking > > through preserved emailboxes; or perhaps someone who printed out all > their > > email still has those printouts). > > > > I feel sadly wary that a lot of our earliest history has been lost > (since we > > didn't use physical memos, which many technical histories depend on for > 'nuts > > and bolts' primary sources) - except for the copy stored in 'meat' CPUs > (who > > will soon start dying off - historians take note, and act now, while you > can). > > > > Noel > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From pugs78 at gmail.com Mon Jun 1 18:41:47 2026 From: pugs78 at gmail.com (Tom Lyon) Date: Mon, 1 Jun 2026 18:41:47 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> Message-ID: Another reason to call them routers was the fact that they ran dynamic routing protocols; very much unlike the typical staticly configured application gateways. 4.2BSD had 'routed' in 1983. (Many struggles with that at Sun) On Mon, Jun 1, 2026 at 6:32?PM vinton cerf via Internet-history < internet-history at elists.isoc.org> wrote: > I had always thought that cisco introduced the term 'router" in 1984. > > v > > > On Mon, Jun 1, 2026 at 8:52?PM Brian E Carpenter via Internet-history < > internet-history at elists.isoc.org> wrote: > > > This is slightly complicated by the IGP/EGP terminology. That persisted > > even when > > "router" became commonplace. I think we can date it to between June 1988 > > and June 1989: > > > > RFC1058 (RIP, June 1988) uses "gateway" exclusively > > > > RFC1105 (the first version of BGP, June 1989) uses "router" exclusively, > > except in the name of the protocol! > > > > Wikipedia says that the p4200 came out in 1986. I couldn't find a manual, > > but its product name was in a May 1988 DoD report: > > > > "Proteon p4200 Gateway > > ... > > The p4200 gateway is a multiprotocol router, supporting (among other > > protocols) TCP/IP." > > > > [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf] > > > > Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) > > > > Regards/Ng? mihi > > Brian > > > > On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: > > > > From: Jack Haverty > > > > > > > In the early Internet, the boxes interconnecting networks were > > called > > > > "gateways". Today they're called "routers". But why the > > change...? > > > > > > > > So we started callig them "routers". Other companies (cisco, > > proteon, > > > > ...) probably had similar experiences in their sales activities. > > > > > > If my memory isn't failing me (it well might be), I can take part of > the > > > blame. > > > > > > I do remember that I was pissed off because everyone and their brother > > > (across the industry generally) called any box that did digital > > > communications between two things a 'gateway'. E.g. a box that did > email > > > forwarding from BITNET to the Internet was called a 'gateway'. I.e. > > 'gateway' > > > was useless as a technical term, because it covered an impossibly wide > > range > > > of functionalities. > > > > > > (I am not sure if the p4200, the first Proteon router product, > > > post-dated the 'gateway' -> 'router' change; I'd have to try and find > an > > > original manual. If it pre-dated, I may have taken Proteon experience > > into > > > account too.) > > > > > > So I campaigned (I think it was me) in the IETF community to come up > > with a > > > term limited to internetwork-level datagra packet switches, and > 'router' > > was > > > picked. > > > > > > > > > I don't know if that change post-dated the creation of the IETF or > not. I > > > remember such large-scale questions (i.e. not within the purview of a > WG, > > > after Phill set up the WG structure) were often discussed on the main > > IETF > > > mailing list, so if we still have the email archive from the start of > > that > > > list, someone can dig into it. > > > > > > I remember that before the IETF existed, there was an email list (I > think > > > hosted at CNRI maybe, although CNRI didn't exist until 1986 - Jon's > > minutes > > > of TCP/IP meetings stop at the end of 1980) where a lot of early TCP > > > internetworking discussions ('TCP internetworking' since there must > have > > been > > > PUP internetworking discussions, too, inside Xerox) happened. Does > anyone > > > remember what it was called? > > > > > > Any technical history of the creation of TCP internetting would > _really_ > > > benefit from having access to that email archive (if it still exists > > > somewhere; if not, maybe it would be possible to re-create it by > picking > > > through preserved emailboxes; or perhaps someone who printed out all > > their > > > email still has those printouts). > > > > > > I feel sadly wary that a lot of our earliest history has been lost > > (since we > > > didn't use physical memos, which many technical histories depend on for > > 'nuts > > > and bolts' primary sources) - except for the copy stored in 'meat' CPUs > > (who > > > will soon start dying off - historians take note, and act now, while > you > > can). > > > > > > Noel > > -- > > Internet-history mailing list > > Internet-history at elists.isoc.org > > https://elists.isoc.org/mailman/listinfo/internet-history > > - > > Unsubscribe: > > > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From gregskinner0 at icloud.com Mon Jun 1 18:54:41 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Mon, 1 Jun 2026 18:54:41 -0700 Subject: [ih] Early Net Number List In-Reply-To: <1573240944.117378.1780353779378@mail.yahoo.com> References: <1856673204.92511.1780352309619.ref@mail.yahoo.com> <1856673204.92511.1780352309619@mail.yahoo.com> <1573240944.117378.1780353779378@mail.yahoo.com> Message-ID: Forwarded for Barbara > From: Barbara Denny > To: Internet-history > Sent: Monday, June 1, 2026 at 03:18:29 PM PDT > Subject: Early Net Number List > > Anyone want a copy of a somewhat early net number list? Unfortunately it is not dated (1983 maybe??). I can't say i have all the entries as it is not alphabetized. The comment says Internet networks are reordered for local efficiency. It starts with srinet and ends with ndre-ring (fits on one page). > > It has been in my garage..... > > barbara From dwaitzman at gmail.com Mon Jun 1 18:55:21 2026 From: dwaitzman at gmail.com (D Waitzman) Date: Mon, 1 Jun 2026 21:55:21 -0400 Subject: [ih] capacity v bandwidth Message-ID: <0459B005-3376-4C3D-8A6F-D0537FEB827A@gmail.com> On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: >> From: Jack Haverty > >> In the early Internet, the boxes interconnecting networks were called >> "gateways". Today they're called "routers". But why the change...? >> >> So we started callig them "routers". Other companies (cisco, proteon, >> ...) probably had similar experiences in their sales activities. I posted my copy of the CMU ECE router source code from around 1986 to github a few years ago at https://github.com/djw1149/cmu-ece-router This code uses a mix of "gateway" and "router": https://github.com/djw1149/cmu-ece-router/blob/master/sys/rcp.c from 07-Mar-84 is the Router Control Protocol module https://github.com/djw1149/cmu-ece-router/blob/master/sys/ipI.c has comment: * 21-Jun-84 Mike Accetta (mja) at Carnegie-Mellon University * Removed support for obsolete "probe" function and changed to * record a single well-known gateway (for systems without EGP); * adjusted to provide the new reporting gateway parameter to * ip_remap() and to initialize the autonomous system number field * of routing entries created for our directly connected networks. -- David Waitzman From dwaitzman at gmail.com Mon Jun 1 18:55:21 2026 From: dwaitzman at gmail.com (D Waitzman) Date: Mon, 1 Jun 2026 21:55:21 -0400 Subject: [ih] capacity v bandwidth Message-ID: <0459B005-3376-4C3D-8A6F-D0537FEB827A@gmail.com> On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: >> From: Jack Haverty > >> In the early Internet, the boxes interconnecting networks were called >> "gateways". Today they're called "routers". But why the change...? >> >> So we started callig them "routers". Other companies (cisco, proteon, >> ...) probably had similar experiences in their sales activities. I posted my copy of the CMU ECE router source code from around 1986 to github a few years ago at https://github.com/djw1149/cmu-ece-router This code uses a mix of "gateway" and "router": https://github.com/djw1149/cmu-ece-router/blob/master/sys/rcp.c from 07-Mar-84 is the Router Control Protocol module https://github.com/djw1149/cmu-ece-router/blob/master/sys/ipI.c has comment: * 21-Jun-84 Mike Accetta (mja) at Carnegie-Mellon University * Removed support for obsolete "probe" function and changed to * record a single well-known gateway (for systems without EGP); * adjusted to provide the new reporting gateway parameter to * ip_remap() and to initialize the autonomous system number field * of routing entries created for our directly connected networks. -- David Waitzman From brian.e.carpenter at gmail.com Mon Jun 1 19:41:59 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Tue, 2 Jun 2026 14:41:59 +1200 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> Message-ID: <24168263-8521-4253-b35a-bf8a2ae11add@gmail.com> Vint, I'm sure you're correct, but again quoting the same 1988 DDN document: "3.8.2. cisco Systems Gateways - PRODUCT-OR-PACKAGE-NAME: cisco Multi-Protocol Gateway Servers DESCRIPTION: The cisco family of gateways are multi-protocol routers linking networks of heterogeneous hosts and - media. All Gateway Servers are fully compliant with RFC 1009, "Requirements for Internet Gateways"." so clearly even Cisco was using both words. RFC 1009 (June 1987) uses "gateway" 154 times and "router" 16 times. It explicitly defines the terms about 180 degrees differently than Noel: " Router A router is a switch that receives data transmission units from input interfaces and, depending on the addresses in those units, routes them to the appropriate output interfaces. There can be routers at different levels of protocol. For example, Interface Message Processors (IMPs) are packet-level routers. Gateway In the Internet documentation generally, and in this document specifically, a gateway is an IP-level router. In the Internet community the term has a long history of this usage [32]." Regards/Ng? mihi Brian On 02-Jun-26 13:32, vinton cerf wrote: > I had always thought that cisco introduced the term 'router" in 1984. > > v > > > On Mon, Jun 1, 2026 at 8:52?PM Brian E Carpenter via Internet-history > wrote: > > This is slightly complicated by the IGP/EGP terminology. That persisted even when > "router" became commonplace. I think we can date it to between June 1988 > and June 1989: > > RFC1058 (RIP, June 1988) uses "gateway" exclusively > > RFC1105 (the first version of BGP, June 1989) uses "router" exclusively, > except in the name of the protocol! > > Wikipedia says that the p4200 came out in 1986. I couldn't find a manual, > but its product name was in a May 1988 DoD report: > > "Proteon p4200 Gateway > ... > The p4200 gateway is a multiprotocol router, supporting (among other protocols) TCP/IP." > > [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf ] > > Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) > > Regards/Ng? mihi > ? ? Brian > > On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: > >? ? ? > From: Jack Haverty > > > >? ? ? > In the early Internet, the boxes interconnecting networks were called > >? ? ? > "gateways". Today they're called "routers".? But why the change...? > >? ? ? > > >? ? ? > So we started callig them "routers". Other companies (cisco, proteon, > >? ? ? > ...) probably had similar experiences in their sales activities. > > > > If my memory isn't failing me (it well might be), I can take part of the > > blame. > > > > I do remember that I was pissed off because everyone and their brother > > (across the industry generally) called any box that did digital > > communications between two things a 'gateway'. E.g. a box that did email > > forwarding from BITNET to the Internet was called a 'gateway'. I.e. 'gateway' > > was useless as a technical term, because it covered an impossibly wide range > > of functionalities. > > > > (I am not sure if the p4200, the first Proteon router product, > > post-dated the 'gateway' -> 'router' change; I'd have to try and find an > > original manual. If it pre-dated, I may have taken Proteon experience into > > account too.) > > > > So I campaigned (I think it was me) in the IETF community to come up with a > > term limited to internetwork-level datagra packet switches, and 'router' was > > picked. > > > > > > I don't know if that change post-dated the creation of the IETF or not. I > > remember such large-scale questions (i.e. not within the purview of a WG, > > after Phill set up the WG structure) were often discussed on the main IETF > > mailing list, so if we still have the email archive from the start of that > > list, someone can dig into it. > > > > I remember that before the IETF existed, there was an email list (I think > > hosted at CNRI maybe, although CNRI didn't exist until 1986 - Jon's minutes > > of TCP/IP meetings stop at the end of 1980) where a lot of early TCP > > internetworking discussions ('TCP internetworking' since there must have been > > PUP internetworking discussions, too, inside Xerox) happened. Does anyone > > remember what it was called? > > > > Any technical history of the creation of TCP internetting would _really_ > > benefit from having access to that email archive (if it still exists > > somewhere; if not, maybe it would be possible to re-create it by picking > > through preserved emailboxes; or perhaps someone who printed out all their > > email still has those printouts). > > > > I feel sadly wary that a lot of our earliest history has been lost (since we > > didn't use physical memos, which many technical histories depend on for 'nuts > > and bolts' primary sources) - except for the copy stored in 'meat' CPUs (who > > will soon start dying off - historians take note, and act now, while you can). > > > >? ? ? ?Noel > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From b_a_denny at yahoo.com Mon Jun 1 20:01:46 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Tue, 2 Jun 2026 03:01:46 +0000 (UTC) Subject: [ih] capacity v bandwidth In-Reply-To: <24168263-8521-4253-b35a-bf8a2ae11add@gmail.com> References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> <24168263-8521-4253-b35a-bf8a2ae11add@gmail.com> Message-ID: <1945056108.178192.1780369306604@mail.yahoo.com> A while ago i unearthed? a pretty early Cisco manual (it only used plastic sheets and a slide binder). If i can can find it again i will see if it says anything on the topic. barbara On Monday, June 1, 2026 at 07:42:14 PM PDT, Brian E Carpenter via Internet-history wrote: Vint, I'm sure you're correct, but again quoting the same 1988 DDN document: "3.8.2. cisco Systems Gateways - PRODUCT-OR-PACKAGE-NAME: cisco Multi-Protocol Gateway Servers DESCRIPTION: The cisco family of gateways are multi-protocol routers linking networks of heterogeneous hosts and - media. All Gateway Servers are fully compliant with RFC 1009, "Requirements for Internet Gateways"." so clearly even Cisco was using both words. RFC 1009 (June 1987) uses "gateway" 154 times and "router" 16 times. It explicitly defines the terms about 180 degrees differently than Noel: "? ? Router? ? ? A router is a switch that receives data transmission ? ? ? ? ? ? ? ? ? units from input interfaces and, depending on the ? ? ? ? ? ? ? ? ? addresses in those units, routes them to the ? ? ? ? ? ? ? ? ? appropriate output interfaces.? There can be routers ? ? ? ? ? ? ? ? ? at different levels of protocol.? For example, ? ? ? ? ? ? ? ? ? Interface Message Processors (IMPs) are packet-level ? ? ? ? ? ? ? ? ? routers. ? ? ? Gateway? ? In the Internet documentation generally, and in this ? ? ? ? ? ? ? ? ? document specifically, a gateway is an IP-level ? ? ? ? ? ? ? ? ? router.? In the Internet community the term has a long ? ? ? ? ? ? ? ? ? history of this usage [32]." Regards/Ng? mihi ? ? Brian On 02-Jun-26 13:32, vinton cerf wrote: > I had always thought that cisco introduced the term 'router" in 1984. > > v > > > On Mon, Jun 1, 2026 at 8:52?PM Brian E Carpenter via Internet-history > wrote: > >? ? This is slightly complicated by the IGP/EGP terminology. That persisted even when >? ? "router" became commonplace. I think we can date it to between June 1988 >? ? and June 1989: > >? ? RFC1058 (RIP, June 1988) uses "gateway" exclusively > >? ? RFC1105 (the first version of BGP, June 1989) uses "router" exclusively, >? ? except in the name of the protocol! > >? ? Wikipedia says that the p4200 came out in 1986. I couldn't find a manual, >? ? but its product name was in a May 1988 DoD report: > >? ? "Proteon p4200 Gateway >? ? ... >? ? The p4200 gateway is a multiprotocol router, supporting (among other protocols) TCP/IP." > >? ? [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf ] > >? ? Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) > >? ? Regards/Ng? mihi >? ? ? ? ? Brian > >? ? On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: >? ? ? >? ? ? > From: Jack Haverty >? ? ? > >? ? ? >? ? ? > In the early Internet, the boxes interconnecting networks were called >? ? ? >? ? ? > "gateways". Today they're called "routers".? But why the change...? >? ? ? >? ? ? > >? ? ? >? ? ? > So we started callig them "routers". Other companies (cisco, proteon, >? ? ? >? ? ? > ...) probably had similar experiences in their sales activities. >? ? ? > >? ? ? > If my memory isn't failing me (it well might be), I can take part of the >? ? ? > blame. >? ? ? > >? ? ? > I do remember that I was pissed off because everyone and their brother >? ? ? > (across the industry generally) called any box that did digital >? ? ? > communications between two things a 'gateway'. E.g. a box that did email >? ? ? > forwarding from BITNET to the Internet was called a 'gateway'. I.e. 'gateway' >? ? ? > was useless as a technical term, because it covered an impossibly wide range >? ? ? > of functionalities. >? ? ? > >? ? ? > (I am not sure if the p4200, the first Proteon router product, >? ? ? > post-dated the 'gateway' -> 'router' change; I'd have to try and find an >? ? ? > original manual. If it pre-dated, I may have taken Proteon experience into >? ? ? > account too.) >? ? ? > >? ? ? > So I campaigned (I think it was me) in the IETF community to come up with a >? ? ? > term limited to internetwork-level datagra packet switches, and 'router' was >? ? ? > picked. >? ? ? > >? ? ? > >? ? ? > I don't know if that change post-dated the creation of the IETF or not. I >? ? ? > remember such large-scale questions (i.e. not within the purview of a WG, >? ? ? > after Phill set up the WG structure) were often discussed on the main IETF >? ? ? > mailing list, so if we still have the email archive from the start of that >? ? ? > list, someone can dig into it. >? ? ? > >? ? ? > I remember that before the IETF existed, there was an email list (I think >? ? ? > hosted at CNRI maybe, although CNRI didn't exist until 1986 - Jon's minutes >? ? ? > of TCP/IP meetings stop at the end of 1980) where a lot of early TCP >? ? ? > internetworking discussions ('TCP internetworking' since there must have been >? ? ? > PUP internetworking discussions, too, inside Xerox) happened. Does anyone >? ? ? > remember what it was called? >? ? ? > >? ? ? > Any technical history of the creation of TCP internetting would _really_ >? ? ? > benefit from having access to that email archive (if it still exists >? ? ? > somewhere; if not, maybe it would be possible to re-create it by picking >? ? ? > through preserved emailboxes; or perhaps someone who printed out all their >? ? ? > email still has those printouts). >? ? ? > >? ? ? > I feel sadly wary that a lot of our earliest history has been lost (since we >? ? ? > didn't use physical memos, which many technical histories depend on for 'nuts >? ? ? > and bolts' primary sources) - except for the copy stored in 'meat' CPUs (who >? ? ? > will soon start dying off - historians take note, and act now, while you can). >? ? ? > >? ? ? >? ? ? ?Noel >? ? -- From vgcerf at gmail.com Mon Jun 1 20:25:33 2026 From: vgcerf at gmail.com (vinton cerf) Date: Mon, 1 Jun 2026 23:25:33 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: <24168263-8521-4253-b35a-bf8a2ae11add@gmail.com> References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> <24168263-8521-4253-b35a-bf8a2ae11add@gmail.com> Message-ID: thanks Brian - Judy Estrin used the term "brouter" for products from Bridge Communications. v On Mon, Jun 1, 2026 at 10:42?PM Brian E Carpenter < brian.e.carpenter at gmail.com> wrote: > Vint, I'm sure you're correct, but again quoting the same 1988 DDN > document: > > "3.8.2. cisco Systems Gateways - > PRODUCT-OR-PACKAGE-NAME: cisco Multi-Protocol Gateway Servers > DESCRIPTION: > The cisco family of gateways are multi-protocol routers linking networks > of heterogeneous hosts and - > media. All Gateway Servers are fully compliant with RFC 1009, > "Requirements for Internet Gateways"." > > so clearly even Cisco was using both words. > > RFC 1009 (June 1987) uses "gateway" 154 times and "router" 16 times. > It explicitly defines the terms about 180 degrees differently than Noel: > > " Router A router is a switch that receives data transmission > units from input interfaces and, depending on the > addresses in those units, routes them to the > appropriate output interfaces. There can be routers > at different levels of protocol. For example, > Interface Message Processors (IMPs) are packet-level > routers. > > Gateway In the Internet documentation generally, and in this > document specifically, a gateway is an IP-level > router. In the Internet community the term has a long > history of this usage [32]." > > Regards/Ng? mihi > Brian > On 02-Jun-26 13:32, vinton cerf wrote: > > I had always thought that cisco introduced the term 'router" in 1984. > > > > v > > > > > > On Mon, Jun 1, 2026 at 8:52?PM Brian E Carpenter via Internet-history < > internet-history at elists.isoc.org > > wrote: > > > > This is slightly complicated by the IGP/EGP terminology. That > persisted even when > > "router" became commonplace. I think we can date it to between June > 1988 > > and June 1989: > > > > RFC1058 (RIP, June 1988) uses "gateway" exclusively > > > > RFC1105 (the first version of BGP, June 1989) uses "router" > exclusively, > > except in the name of the protocol! > > > > Wikipedia says that the p4200 came out in 1986. I couldn't find a > manual, > > but its product name was in a May 1988 DoD report: > > > > "Proteon p4200 Gateway > > ... > > The p4200 gateway is a multiprotocol router, supporting (among other > protocols) TCP/IP." > > > > [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf < > https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf>] > > > > Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) > > > > Regards/Ng? mihi > > Brian > > > > On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: > > > > From: Jack Haverty > > > > > > > In the early Internet, the boxes interconnecting networks > were called > > > > "gateways". Today they're called "routers". But why the > change...? > > > > > > > > So we started callig them "routers". Other companies > (cisco, proteon, > > > > ...) probably had similar experiences in their sales > activities. > > > > > > If my memory isn't failing me (it well might be), I can take part > of the > > > blame. > > > > > > I do remember that I was pissed off because everyone and their > brother > > > (across the industry generally) called any box that did digital > > > communications between two things a 'gateway'. E.g. a box that > did email > > > forwarding from BITNET to the Internet was called a 'gateway'. > I.e. 'gateway' > > > was useless as a technical term, because it covered an impossibly > wide range > > > of functionalities. > > > > > > (I am not sure if the p4200, the first Proteon router product, > > > post-dated the 'gateway' -> 'router' change; I'd have to try and > find an > > > original manual. If it pre-dated, I may have taken Proteon > experience into > > > account too.) > > > > > > So I campaigned (I think it was me) in the IETF community to come > up with a > > > term limited to internetwork-level datagra packet switches, and > 'router' was > > > picked. > > > > > > > > > I don't know if that change post-dated the creation of the IETF > or not. I > > > remember such large-scale questions (i.e. not within the purview > of a WG, > > > after Phill set up the WG structure) were often discussed on the > main IETF > > > mailing list, so if we still have the email archive from the > start of that > > > list, someone can dig into it. > > > > > > I remember that before the IETF existed, there was an email list > (I think > > > hosted at CNRI maybe, although CNRI didn't exist until 1986 - > Jon's minutes > > > of TCP/IP meetings stop at the end of 1980) where a lot of early > TCP > > > internetworking discussions ('TCP internetworking' since there > must have been > > > PUP internetworking discussions, too, inside Xerox) happened. > Does anyone > > > remember what it was called? > > > > > > Any technical history of the creation of TCP internetting would > _really_ > > > benefit from having access to that email archive (if it still > exists > > > somewhere; if not, maybe it would be possible to re-create it by > picking > > > through preserved emailboxes; or perhaps someone who printed out > all their > > > email still has those printouts). > > > > > > I feel sadly wary that a lot of our earliest history has been > lost (since we > > > didn't use physical memos, which many technical histories depend > on for 'nuts > > > and bolts' primary sources) - except for the copy stored in > 'meat' CPUs (who > > > will soon start dying off - historians take note, and act now, > while you can). > > > > > > Noel > > -- > > Internet-history mailing list > > Internet-history at elists.isoc.org Internet-history at elists.isoc.org> > > https://elists.isoc.org/mailman/listinfo/internet-history < > https://elists.isoc.org/mailman/listinfo/internet-history> > > - > > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > < > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > > > > From geoff at iconia.com Mon Jun 1 20:41:22 2026 From: geoff at iconia.com (the keyboard of geoff goodfellow) Date: Mon, 1 Jun 2026 20:41:22 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> <24168263-8521-4253-b35a-bf8a2ae11add@gmail.com> Message-ID: would the very first "gateway" perhaps be the Menehune application level "gateway" between the ALOHANET that "became operational in June 1971, providing the first public demonstration of a wireless packet data network" connected to the ARPANET? https://en.wikipedia.org/wiki/ALOHAnet yours truly spent some summer "vacations" in Hawaii in the '70s and was given use of an ALOHANET ICU to connect to a TI Silent 700 Thermal print out terminal yours truly lugged over there and fondly recalls using the HAWAII-ALOHA Menehune "gateway" to telnet connect back to SRI. sadly yours truly no longer has the photo's of all that.. :( g On Mon, Jun 1, 2026 at 8:25?PM vinton cerf via Internet-history < internet-history at elists.isoc.org> wrote: > thanks Brian - Judy Estrin used the term "brouter" for products from Bridge > Communications. > > v > > > On Mon, Jun 1, 2026 at 10:42?PM Brian E Carpenter < > brian.e.carpenter at gmail.com> wrote: > > > Vint, I'm sure you're correct, but again quoting the same 1988 DDN > > document: > > > > "3.8.2. cisco Systems Gateways - > > PRODUCT-OR-PACKAGE-NAME: cisco Multi-Protocol Gateway Servers > > DESCRIPTION: > > The cisco family of gateways are multi-protocol routers linking networks > > of heterogeneous hosts and - > > media. All Gateway Servers are fully compliant with RFC 1009, > > "Requirements for Internet Gateways"." > > > > so clearly even Cisco was using both words. > > > > RFC 1009 (June 1987) uses "gateway" 154 times and "router" 16 times. > > It explicitly defines the terms about 180 degrees differently than Noel: > > > > " Router A router is a switch that receives data transmission > > units from input interfaces and, depending on the > > addresses in those units, routes them to the > > appropriate output interfaces. There can be routers > > at different levels of protocol. For example, > > Interface Message Processors (IMPs) are packet-level > > routers. > > > > Gateway In the Internet documentation generally, and in this > > document specifically, a gateway is an IP-level > > router. In the Internet community the term has a long > > history of this usage [32]." > > > > Regards/Ng? mihi > > Brian > > On 02-Jun-26 13:32, vinton cerf wrote: > > > I had always thought that cisco introduced the term 'router" in 1984. > > > > > > v > > > > > > > > > On Mon, Jun 1, 2026 at 8:52?PM Brian E Carpenter via Internet-history < > > internet-history at elists.isoc.org internet-history at elists.isoc.org>> > > wrote: > > > > > > This is slightly complicated by the IGP/EGP terminology. That > > persisted even when > > > "router" became commonplace. I think we can date it to between June > > 1988 > > > and June 1989: > > > > > > RFC1058 (RIP, June 1988) uses "gateway" exclusively > > > > > > RFC1105 (the first version of BGP, June 1989) uses "router" > > exclusively, > > > except in the name of the protocol! > > > > > > Wikipedia says that the p4200 came out in 1986. I couldn't find a > > manual, > > > but its product name was in a May 1988 DoD report: > > > > > > "Proteon p4200 Gateway > > > ... > > > The p4200 gateway is a multiprotocol router, supporting (among > other > > protocols) TCP/IP." > > > > > > [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf < > > https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf>] > > > > > > Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) > > > > > > Regards/Ng? mihi > > > Brian > > > > > > On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: > > > > > From: Jack Haverty > > > > > > > > > In the early Internet, the boxes interconnecting networks > > were called > > > > > "gateways". Today they're called "routers". But why the > > change...? > > > > > > > > > > So we started callig them "routers". Other companies > > (cisco, proteon, > > > > > ...) probably had similar experiences in their sales > > activities. > > > > > > > > If my memory isn't failing me (it well might be), I can take > part > > of the > > > > blame. > > > > > > > > I do remember that I was pissed off because everyone and their > > brother > > > > (across the industry generally) called any box that did digital > > > > communications between two things a 'gateway'. E.g. a box that > > did email > > > > forwarding from BITNET to the Internet was called a 'gateway'. > > I.e. 'gateway' > > > > was useless as a technical term, because it covered an > impossibly > > wide range > > > > of functionalities. > > > > > > > > (I am not sure if the p4200, the first Proteon router product, > > > > post-dated the 'gateway' -> 'router' change; I'd have to try and > > find an > > > > original manual. If it pre-dated, I may have taken Proteon > > experience into > > > > account too.) > > > > > > > > So I campaigned (I think it was me) in the IETF community to > come > > up with a > > > > term limited to internetwork-level datagra packet switches, and > > 'router' was > > > > picked. > > > > > > > > > > > > I don't know if that change post-dated the creation of the IETF > > or not. I > > > > remember such large-scale questions (i.e. not within the purview > > of a WG, > > > > after Phill set up the WG structure) were often discussed on the > > main IETF > > > > mailing list, so if we still have the email archive from the > > start of that > > > > list, someone can dig into it. > > > > > > > > I remember that before the IETF existed, there was an email list > > (I think > > > > hosted at CNRI maybe, although CNRI didn't exist until 1986 - > > Jon's minutes > > > > of TCP/IP meetings stop at the end of 1980) where a lot of early > > TCP > > > > internetworking discussions ('TCP internetworking' since there > > must have been > > > > PUP internetworking discussions, too, inside Xerox) happened. > > Does anyone > > > > remember what it was called? > > > > > > > > Any technical history of the creation of TCP internetting would > > _really_ > > > > benefit from having access to that email archive (if it still > > exists > > > > somewhere; if not, maybe it would be possible to re-create it by > > picking > > > > through preserved emailboxes; or perhaps someone who printed out > > all their > > > > email still has those printouts). > > > > > > > > I feel sadly wary that a lot of our earliest history has been > > lost (since we > > > > didn't use physical memos, which many technical histories depend > > on for 'nuts > > > > and bolts' primary sources) - except for the copy stored in > > 'meat' CPUs (who > > > > will soon start dying off - historians take note, and act now, > > while you can). > > > > > > > > Noel > > > -- > > > Internet-history mailing list > > > Internet-history at elists.isoc.org > Internet-history at elists.isoc.org> > > > https://elists.isoc.org/mailman/listinfo/internet-history < > > https://elists.isoc.org/mailman/listinfo/internet-history> > > > - > > > Unsubscribe: > > > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > < > > > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > > > > > > > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > -- Geoff.Goodfellow at iconia.com living as The Truth is True From gregskinner0 at icloud.com Tue Jun 2 00:34:36 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Tue, 2 Jun 2026 00:34:36 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> Message-ID: On Jun 1, 2026, at 6:41?PM, Tom Lyon via Internet-history wrote: > > Another reason to call them routers was the fact that they ran dynamic > routing protocols; very much unlike the typical staticly configured > application gateways. > > 4.2BSD had 'routed' in 1983. (Many struggles with that at Sun) > In ISI research report ISI/RR-84-145 , dated October 1984, ?router? appears in the egpup (EGP user process) logs to designate the EGP neighbor. --gregbo From vgcerf at gmail.com Tue Jun 2 02:41:39 2026 From: vgcerf at gmail.com (vinton cerf) Date: Tue, 2 Jun 2026 05:41:39 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260601234820.B548A18C074@mercury.lcs.mit.edu> <0cf7ca9e-5383-459e-a459-8b550753f9c3@gmail.com> <24168263-8521-4253-b35a-bf8a2ae11add@gmail.com> Message-ID: I think of menehune as a wireless LAN, so it's more an access point than a gateway. On Mon, Jun 1, 2026, 23:42 the keyboard of geoff goodfellow < geoff at iconia.com> wrote: > would the very first "gateway" perhaps be the Menehune application level > "gateway" between the ALOHANET that "became operational in June 1971, > providing the first public demonstration of a wireless packet data network" > connected to the ARPANET? > https://en.wikipedia.org/wiki/ALOHAnet > > yours truly spent some summer "vacations" in Hawaii in the '70s and was > given use of an ALOHANET ICU to connect to a TI Silent 700 Thermal print > out terminal yours truly lugged over there and fondly recalls using the > HAWAII-ALOHA Menehune "gateway" to telnet connect back to SRI. sadly yours > truly no longer has the photo's of all that.. :( > > g > > On Mon, Jun 1, 2026 at 8:25?PM vinton cerf via Internet-history < > internet-history at elists.isoc.org> wrote: > >> thanks Brian - Judy Estrin used the term "brouter" for products from >> Bridge >> Communications. >> >> v >> >> >> On Mon, Jun 1, 2026 at 10:42?PM Brian E Carpenter < >> brian.e.carpenter at gmail.com> wrote: >> >> > Vint, I'm sure you're correct, but again quoting the same 1988 DDN >> > document: >> > >> > "3.8.2. cisco Systems Gateways - >> > PRODUCT-OR-PACKAGE-NAME: cisco Multi-Protocol Gateway Servers >> > DESCRIPTION: >> > The cisco family of gateways are multi-protocol routers linking networks >> > of heterogeneous hosts and - >> > media. All Gateway Servers are fully compliant with RFC 1009, >> > "Requirements for Internet Gateways"." >> > >> > so clearly even Cisco was using both words. >> > >> > RFC 1009 (June 1987) uses "gateway" 154 times and "router" 16 times. >> > It explicitly defines the terms about 180 degrees differently than Noel: >> > >> > " Router A router is a switch that receives data transmission >> > units from input interfaces and, depending on the >> > addresses in those units, routes them to the >> > appropriate output interfaces. There can be routers >> > at different levels of protocol. For example, >> > Interface Message Processors (IMPs) are packet-level >> > routers. >> > >> > Gateway In the Internet documentation generally, and in this >> > document specifically, a gateway is an IP-level >> > router. In the Internet community the term has a >> long >> > history of this usage [32]." >> > >> > Regards/Ng? mihi >> > Brian >> > On 02-Jun-26 13:32, vinton cerf wrote: >> > > I had always thought that cisco introduced the term 'router" in 1984. >> > > >> > > v >> > > >> > > >> > > On Mon, Jun 1, 2026 at 8:52?PM Brian E Carpenter via Internet-history >> < >> > internet-history at elists.isoc.org > internet-history at elists.isoc.org>> >> > wrote: >> > > >> > > This is slightly complicated by the IGP/EGP terminology. That >> > persisted even when >> > > "router" became commonplace. I think we can date it to between >> June >> > 1988 >> > > and June 1989: >> > > >> > > RFC1058 (RIP, June 1988) uses "gateway" exclusively >> > > >> > > RFC1105 (the first version of BGP, June 1989) uses "router" >> > exclusively, >> > > except in the name of the protocol! >> > > >> > > Wikipedia says that the p4200 came out in 1986. I couldn't find a >> > manual, >> > > but its product name was in a May 1988 DoD report: >> > > >> > > "Proteon p4200 Gateway >> > > ... >> > > The p4200 gateway is a multiprotocol router, supporting (among >> other >> > protocols) TCP/IP." >> > > >> > > [https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf < >> > https://apps.dtic.mil/sti/tr/pdf/ADA192186.pdf>] >> > > >> > > Also see: RFC1208 "A Glossary of Networking Terms" (March 1991) >> > > >> > > Regards/Ng? mihi >> > > Brian >> > > >> > > On 02-Jun-26 11:48, Noel Chiappa via Internet-history wrote: >> > > > > From: Jack Haverty >> > > > >> > > > > In the early Internet, the boxes interconnecting >> networks >> > were called >> > > > > "gateways". Today they're called "routers". But why the >> > change...? >> > > > > >> > > > > So we started callig them "routers". Other companies >> > (cisco, proteon, >> > > > > ...) probably had similar experiences in their sales >> > activities. >> > > > >> > > > If my memory isn't failing me (it well might be), I can take >> part >> > of the >> > > > blame. >> > > > >> > > > I do remember that I was pissed off because everyone and their >> > brother >> > > > (across the industry generally) called any box that did digital >> > > > communications between two things a 'gateway'. E.g. a box that >> > did email >> > > > forwarding from BITNET to the Internet was called a 'gateway'. >> > I.e. 'gateway' >> > > > was useless as a technical term, because it covered an >> impossibly >> > wide range >> > > > of functionalities. >> > > > >> > > > (I am not sure if the p4200, the first Proteon router product, >> > > > post-dated the 'gateway' -> 'router' change; I'd have to try >> and >> > find an >> > > > original manual. If it pre-dated, I may have taken Proteon >> > experience into >> > > > account too.) >> > > > >> > > > So I campaigned (I think it was me) in the IETF community to >> come >> > up with a >> > > > term limited to internetwork-level datagra packet switches, and >> > 'router' was >> > > > picked. >> > > > >> > > > >> > > > I don't know if that change post-dated the creation of the IETF >> > or not. I >> > > > remember such large-scale questions (i.e. not within the >> purview >> > of a WG, >> > > > after Phill set up the WG structure) were often discussed on >> the >> > main IETF >> > > > mailing list, so if we still have the email archive from the >> > start of that >> > > > list, someone can dig into it. >> > > > >> > > > I remember that before the IETF existed, there was an email >> list >> > (I think >> > > > hosted at CNRI maybe, although CNRI didn't exist until 1986 - >> > Jon's minutes >> > > > of TCP/IP meetings stop at the end of 1980) where a lot of >> early >> > TCP >> > > > internetworking discussions ('TCP internetworking' since there >> > must have been >> > > > PUP internetworking discussions, too, inside Xerox) happened. >> > Does anyone >> > > > remember what it was called? >> > > > >> > > > Any technical history of the creation of TCP internetting would >> > _really_ >> > > > benefit from having access to that email archive (if it still >> > exists >> > > > somewhere; if not, maybe it would be possible to re-create it >> by >> > picking >> > > > through preserved emailboxes; or perhaps someone who printed >> out >> > all their >> > > > email still has those printouts). >> > > > >> > > > I feel sadly wary that a lot of our earliest history has been >> > lost (since we >> > > > didn't use physical memos, which many technical histories >> depend >> > on for 'nuts >> > > > and bolts' primary sources) - except for the copy stored in >> > 'meat' CPUs (who >> > > > will soon start dying off - historians take note, and act now, >> > while you can). >> > > > >> > > > Noel >> > > -- >> > > Internet-history mailing list >> > > Internet-history at elists.isoc.org > > Internet-history at elists.isoc.org> >> > > https://elists.isoc.org/mailman/listinfo/internet-history < >> > https://elists.isoc.org/mailman/listinfo/internet-history> >> > > - >> > > Unsubscribe: >> > >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > < >> > >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > > >> > > >> > >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> >> > > -- > Geoff.Goodfellow at iconia.com > living as The Truth is True > > From jnc at mercury.lcs.mit.edu Tue Jun 2 03:28:04 2026 From: jnc at mercury.lcs.mit.edu (Noel Chiappa) Date: Tue, 2 Jun 2026 06:28:04 -0400 (EDT) Subject: [ih] capacity v bandwidth Message-ID: <20260602102804.1569518C074@mercury.lcs.mit.edu> > From: Brian E Carpenter > I think we can date it to between June 1988 and June 1989: > RFC1058 (RIP, June 1988) uses "gateway" exclusively > RFC1105 (the first version of BGP, June 1989) uses "router" > exclusively, except in the name of the protocol! Excellent detective work on the timing! Given the date, the discussion about doing it should have happened on the IETF list (the first IETF meeting was in January, 1986, per: https://www.ietf.org/meeting/past/ so I'm sure the IETF mailing list was in use by June, 1988). Does anyone know how far back the IETF mailing lis archives go back (and where they are - unless those are _also_ lost, unless someone starts an effort to re-create them - a worthwhile project, as internetting is the biggest advance in information technology since Gutenberg and the telegraph)? Unless, of course, it was done as the result of in-person discussion at an IETF meeting (which, if done in a hallway, might not have left a trace in a Proceedings). I would not make any guesses (based on various clues, e.g. where its use turns up) about _why_ it was done, unless we can locate records of the discussion to change the term. (There's a reason that, for professional historians, contemporaneous written records are the 'gold standard'. All other sources - especially human memory - are very inferior.) (A good example of this in action can be found here: http://www.chiappa.net/~jnc/nontech/tmlotus.html although the finale there was technically of course an old photo, not "written records" - although old photos are even better, as humans can make errors in writing things down.) Noel From b_a_denny at yahoo.com Tue Jun 2 08:18:50 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Tue, 2 Jun 2026 15:18:50 +0000 (UTC) Subject: [ih] capacity v bandwidth In-Reply-To: <20260602102804.1569518C074@mercury.lcs.mit.edu> References: <20260602102804.1569518C074@mercury.lcs.mit.edu> Message-ID: <1286122081.338167.1780413530808@mail.yahoo.com> I doubt it was discussed in a hallway.? To me the early meetings were more like a DARPA meeting which covered more than one project.? People presented their work/ideas and people would discuss it, a very different feeling and style then later. Everyone was in one room as far as I knew. The number of people attending was not that high. I was at the fourth meeting at SRI. Granted I did give a demo of RP and that was a small group.?? barbara On Tuesday, June 2, 2026 at 03:28:14 AM PDT, Noel Chiappa via Internet-history wrote: ? ? > From: Brian E Carpenter ? ? > I think we can date it to between June 1988 and June 1989: ? ? > RFC1058 (RIP, June 1988) uses "gateway" exclusively ? ? > RFC1105 (the first version of BGP, June 1989) uses "router" ? ? > exclusively, except in the name of the protocol! Excellent detective work on the timing! Given the date, the discussion about doing it should have happened on the IETF list (the first IETF meeting was in January, 1986, per: ? https://www.ietf.org/meeting/past/ so I'm sure the IETF mailing list was in use by June, 1988). Does anyone know how far back the IETF mailing lis archives go back (and where they are - unless those are _also_ lost, unless someone starts an effort to re-create them - a worthwhile project, as internetting is the biggest advance in information technology since Gutenberg and the telegraph)? Unless, of course, it was done as the result of in-person discussion at an IETF meeting (which, if done in a hallway, might not have left a trace in a Proceedings). I would not make any guesses (based on various clues, e.g. where its use turns up) about _why_ it was done, unless we can locate records of the discussion to change the term. (There's a reason that, for professional historians, contemporaneous written records are the 'gold standard'. All other sources - especially human memory - are very inferior.) (A good example of this in action can be found here: ??http://www.chiappa.net/~jnc/nontech/tmlotus.html although the finale there was technically of course an old photo, not "written records" - although old photos are even better, as humans can make errors in writing things down.) ??? Noel -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From dwaitzman at gmail.com Tue Jun 2 08:53:10 2026 From: dwaitzman at gmail.com (D Waitzman) Date: Tue, 2 Jun 2026 11:53:10 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: References: Message-ID: Regarding the history subtopic: https://www.iab.org/about/history/ has: > The origin of today?s IAB lies in the Internet Configuration Control Board (ICCB), which was created in 1979 by Vint Cerf, at that time program manager at DARPA, to advise him on technical issues. The ICCB was chaired by David Clark, MIT. > > In September 1984, after the ICCB meeting held at RSRE in Malvern, UK, the ICCB was disbanded and replaced by the Internet Advisory Board (IAB). This change was initiated by Dave Clark and Barry Leiner, who had taken over management of the Internet research program at DARPA. The IAB consisted of the chairs of the newly-formed research task forces and Jon Postel (ISI), as RFC editor and ?protocol czar?. The first set of chairs of the task forces were the members of the ICCB. The IAB was chaired by Dave Clark. I thought it was called the ICB, not ICCB. I attended a meeting in The Hague at NATO's STC around 1991. MAJ (or COL by then?) Pullen was funding us. Noel Chiappa wrote: > There's a reason that, for professional historians, contemporaneous written records are the 'gold standard'. All other sources - especially human memory - are very inferior. Circa 1989 on a Friday afternoon, Steve Storch asked me to fly to California for a Monday ICB or IETF Host Requirements meeting to take notes. Fortunately, BBN had a capable onsite travel agent in our building (10 Moulton) who quickly booked me tickets (pre-web this was a big deal). I emailed out the next on Tuesday to Bob Braden, who emailed them out to the team with a subject line like "Shazam!" since the notes usually took longer to be written up and distributed. (I was amused by the cost of a note taker for that contract.) Those certainly meet Noel's gold standard, but I cannot find that email. -- David Waitzman From mfidelman at protocoltechnologiesgroup.com Tue Jun 2 10:08:52 2026 From: mfidelman at protocoltechnologiesgroup.com (Miles Fidelman) Date: Tue, 2 Jun 2026 17:08:52 +0000 Subject: [ih] A small story of IMP #1 and the UCLA Computer Club In-Reply-To: <0CCE4623-49CE-4BCF-AEC6-6F09178FA36D@comcast.net> References: <0CCE4623-49CE-4BCF-AEC6-6F09178FA36D@comcast.net> Message-ID: Reminds me of the time Bill Silver (then a roommate) answered a question on one of John Donovan's 6.251 tests: Q: What happens when the bit bucket overflows? Bill's Answer: "The bits spill out all over the floor." Miles ________________________________ From: Internet-history on behalf of John Day via Internet-history Sent: Monday, June 1, 2026 7:28 PM To: Vint Cerf ; Vint Cerf via Internet-history Subject: Re: [ih] A small story of IMP #1 and the UCLA Computer Club We once convinced a poor undergrad that every night, the operator had to empty the bit-bucket from all of the left and right shifts during the day. To convince the poor student, they were shown the card chad from a keypunch no one used. > On Jun 1, 2026, at 17:37, Vint Cerf via Internet-history wrote: > > While we are in anecdote mode, the Sigma-7, like other processors, could be > used to toggle a flip-flop at various rates to produce sound. This was > usually accomplished by putting the machine into a loop with a certain > repetition rate. One of the senior graduate students was into computer > music. One day, I brought a friend to see the Sigma-7 and the IMP. When we > arrived, the sr. grad. had a set of tuning forks and was adjusting the loop > lengths to match the sounds of the tuning forks. As we approached, I > cautioned my guest to be quiet. "He's tuning the computer, I said...." > > v > > > On Mon, Jun 1, 2026 at 5:21?PM Andrew G. Malis via Internet-history < > internet-history at elists.isoc.org> wrote: > >> Craig, >> >> I did my share of Pluribus coding and debugging. The assembly code was >> really no worse than any other once you got the hang of it, it was actually >> somewhat similar to PDP-11 assembler. It was a multiprocessor machine, and >> if any of the processors crashed, the remaining processors would notice, >> send a fresh code reload to the failed processor, and restart it. As a >> result, it was a pain to try to debug issues in the field, although crashes >> would produce dumps for later analysis. The easiest way to debug the code >> in the lab was to shut down all but one processor. We were able to shut >> down (and keep down) individual processors by using the front panel >> switches to write "FADE" into a particular location. We could later reboot >> the FADEd processor(s) manually. >> >> Regarding physically stopping a Pluribus by bullets or hand grenade, I >> certainly believe a hand grenade would damage all of the processors enough >> to bring them all down. One well-placed bullet in each processor would >> probably do it as well, but you would have to do it carefully, not just >> fire willy-nilly at the collection of processors, you could easily miss at >> least one of the processors that way. >> >> Cheers, >> Andy >> >> >> >> On Mon, Jun 1, 2026 at 3:40?PM Craig Partridge via Internet-history < >> internet-history at elists.isoc.org> wrote: >> >>> By the time I joined BBN, the story was that the person shooting the >>> Pluribus had been forced to write programs for it. It had a BBN-wide >>> reputation for not being programmer-friendly (which is saying something, >> as >>> BBN periodically generated computing platforms which were painful to >>> program -- such as the C70 with 10-bit bytes). >>> >>> Craig >>> >>> On Mon, Jun 1, 2026 at 12:41?PM Jack Haverty via Internet-history < >>> internet-history at elists.isoc.org> wrote: >>> >>>> Another IMP story... >>>> >>>> The "Pluribus" computer was a BBN creation that included highly >>>> redundant hardware, with multiple everythings. The idea was that no >>>> matter what failed, the system would keep running while repairs were >>>> made. Of course a Pluribus could be used as an IMP, which was popular >>>> in some government installations. One day at BBN, I heard this story. >>>> Wasn't there myself, but I can believe it. >>>> >>>> A Pluribus IMP was being decommissioned at some government site. They >>>> happened to have a variety of military stuff around. So someone >>>> decided to see if the Pluribus IMP was as reliable as it was touted to >>> be. >>>> >>>> The IMP was set up, still running. Someone got a rifle (M-16?) and >>>> started shooting at the IMP. Really. Sadly I don't recall the number, >>>> but the IMP survived an amazing number of direct hits at point blank >>>> range, and still kept passing traffic. >>>> >>>> Where did those folks in the Computer Club go after leaving UCLA...? >>>> Any of them in ROTC? >>>> >>>> /Jack >>>> >>>> On 6/1/26 11:22, Karl Auerbach via Internet-history wrote: >>>>> This is a trivial, and quite irreverent, bit of Internet history.... >>>>> >>>>> I, like several others, were members of the UCLA computer club during >>>>> the late 1960's. The club's office was in Boelter Hall - not far >> from >>>>> the room that held IMP #1 (and the Sigma computer - along with its >>>>> "Sigma EXecutive" documentation, aka "SEX Manuals". My project's >>>>> computer, an IBM 7094 - with a true memory leak [the core memory was >>>>> oil cooled, and that oil leaked] - was in the next room over and we >>>>> could hear the squeals from the AM radio caused by the RF noise from >>>>> the IMP and the Sigma.) >>>>> >>>>> Anyway, folks in the Computer Club - especially Mark Kampe - kinda >>>>> like to pull pranks. For instance, we would drop things from the top >>>>> of Boelter Hall (9 floors up) to see what would happen. The landing >>>>> zone was the collection of crunched and bent automobiles resulting >>>>> from the early crash-tests of my group, the Institute of Traffic and >>>>> Traffic Engineering. We dropped everything from frozen superballs to >>>>> a lead container used to hold/transport radioactive materials [it was >>>>> empty]. There was also a feisty ice-cream vending machine in the >>>>> hallway that once-too-often failed to deliver the paid-for frozen >>>>> treat - so someone in the club unplugged the machine for a few hours, >>>>> everything inside melted, and then plugged it back in, re-frezzing >> the >>>>> leaking drippy mess. That was not nice, but it was - here's a >>>>> terrible pun - that vending machine received its just desserts. >>>>> >>>>> Anyway, back to Internet History... >>>>> >>>>> IMP #1 had the rough appearance of an armor plated refrigerator, with >>>>> lifting lugs on the top. The machine was "ruggedized". >>>>> >>>>> That word, ruggedized, was like honey to ants - it seriously caught >>>>> our attention. So we (I think Mark K. in particular) asked "Is it >>>>> rugged enough to survive a drop from the top of Boelter Hall?". >>>>> >>>>> So our imaginations lit up with images of us grabbing IMP #1, hauling >>>>> it up to the roof and dropping it into the crashed cars nine floors >>>>> below. >>>>> >>>>> Obviously, prudence and sanity - and perhaps even some, probably >>>>> reluctant, respect for law - prevailed. We never did get beyond the >>>>> "what if we did this" stage. >>>>> >>>>> (But a couple of years later some of us migrated from UCLA to SDC in >>>>> Santa Monica. At SDC we had an extremely awful HP 2000 minicomputer >>>>> that, if I remember properly, did actually suffer such a fate as it >>>>> was dropped it from the roof of the Q7A building - three stories tall >>>>> - onto the parking lot - a fate that all of us applauded. [It was >>>>> truly a terrible machine with an even worse operating system.] Some >>>>> of use, years later, moved onto the Interop show nets were we >> sometime >>>>> had to practice the delicate art of percussive maintenance.) >>>>> >>>>> --karl-- >>>>> >>>> >>>> -- >>>> Internet-history mailing list >>>> Internet-history at elists.isoc.org >>>> https://elists.isoc.org/mailman/listinfo/internet-history >>>> - >>>> Unsubscribe: >>>> >>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >>>> >>> >>> >>> -- >>> ***** >>> Craig Partridge's email account for professional society activities and >>> mailing lists. >>> -- >>> Internet-history mailing list >>> Internet-history at elists.isoc.org >>> https://elists.isoc.org/mailman/listinfo/internet-history >>> - >>> Unsubscribe: >>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >>> >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > > > -- > After July 1, Please send any postal/overnight deliveries to: > Vint Cerf > 1435 Woodhurst Blvd > McLean, VA 22102 > > Until July 1, send postal/overnight deliveries to: > Google, LLC > 1900 Reston Metro Plaza, 16th Floor > Reston, VA 20190 > +1 650-224-2788 > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From jnc at mercury.lcs.mit.edu Tue Jun 2 10:28:28 2026 From: jnc at mercury.lcs.mit.edu (Noel Chiappa) Date: Tue, 2 Jun 2026 13:28:28 -0400 (EDT) Subject: [ih] capacity v bandwidth Message-ID: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> > From: Barbara Denny > I doubt it was discussed in a hallway. To me the early meetings were > more like a DARPA meeting ... Everyone was in one room as far as I knew. > The number of people attending was not that high. IETF meetings very quickly became fairly sizeable. From the meeting list I linked to earlier: 1st IETF 21 Attendees January, 1986; San Diego 2nd IETF 21 Attendees April, 1986; Aberdeen 3rd IETF 18 Attendees July, 1986; Ann Arbor 4th IETF 35 Attendees October, 1986; Menlo Park 5th IETF 35 Attendees February, 1987; Moffett Field 6th IETF 88 Attendees April, 1987; Boston 7th IETF 101 Attendees July, 1987; McLean 8th IETF 56 Attendees November, 1987; Boulder 9th IETF 82 Attendees March, 1988; San Diego 10th IETF 112 Attendees June, 1988; Annapolis 11th IETF 114 Attendees October, 1988; Ann Arbor 12th IETF 120 Attendees January, 1989; Austin 13th IETF 114 Attendees April, 1989; Cocoa Beach With the timeframe that Brian deduced for the gateway/router change - between June 1988 and June 1989 - the initial growth had pretty clearly taken off. (The 7th, in McLean, was almost certainly the one where I remember Phill, Dan Lynch and I sitting in a bar at the end of the day chanting 'It's time to get real'; no doubt it being the first one with more than 100 attendees contributedto that thinking.) Noel From jack at 3kitty.org Tue Jun 2 10:30:48 2026 From: jack at 3kitty.org (Jack Haverty) Date: Tue, 2 Jun 2026 10:30:48 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: References: Message-ID: <4caa86b9-253e-48dd-bae6-863952a7c0e0@3kitty.org> Not quite what I remember. The ICCB was the Internet Configuration Control Board.? Bob Kahn called it "Vint's Cabinet".? ?The name was chosen explicitly to make it sound bureaucratic and boring.? ?Otherwise everyone would have wanted to be on it, and a big problem at the time (late 1970s) was the size of the various "internet meetings".? Dave Clark was a member of the ICCB, but Vint was clearly the Chair.? I don't recall a single meeting not run by Vint. The ICB was the International Collaboration Board (or something like that).? There was overlap (e.g., Vint) in the ICB and ICCB memberships.? The ICCB focussed on planning and execution of the steps needed to target a "January 1983" Internet capable of handling "heavy load".? ?The ICB focussed on issues specific to trans-national connectivity, e.g., between the US and Europe, and had members such as John Laws (RSRE), Horst Klaussen (DFVLR), Peter Kirstein (UCL), and Paal Spilling (NDRE).? May have been others too.? Someone from CNUCE (Italy) was probably involved but I never met him or her. The earliest ICCB meeting I can find so far in my notes was on September 21, 1981.? I think it may have been the first meeting of the ICCB, based on what's in the notes I took.? I recall at some point around then getting an email from Vint asking if I could come a day early to the next Internet Meeting, which turned out to be the first meeting of the new ICCB.? I'm not sure, but think it was that September 1981 meeting.? ?My notes don't have an attendee list, but IIRC the ICCB meetings included Jim Mathis, Dave Clark, Jon Postel, Dave Mills, Bob Braden, Ed Cain, Ray McFarland, myself (Jack Haverty), and of course Vint. Meetings of the ICCB were usually held the day before a scheduled Internet Meeting.? We were typically considering what had to happen near-term for a January 1983 target of an Internet capable of handling "heavy load", as well as long-term issues that needed more research.? The 1983 target was clearly associated with the intent to replace NCP on the ARPANET with TCP, and the adoption of TCP as a DoD Standard. It was less than two years away -- not a lot of time to get everything in place.? ?There was a clear conflict between "Research" and "Operations".? Research seeks to study and try new ideas; Operations seeks to get something in place and working as a service.? The Internet needed to do both. After an ICCB meeting we all returned to our regular activities and did whatever we could to get the near-term actions to happen and the long-term issues to get as much attention as possible.? The prioritization helped also for Vint to control the Internet Meeting which immediately followed.? ?"Herding Cats" comes to mind.... I was on the ICCB until Vint's departure from ARPA and Barry Leiner reorganized the ICCB to become the IAB.? ?There had been a major re-organization at BBN in July 1983 and I was no longer to be involved in the "research" contracts. I can't remember if this list accepts attachments -- so I've put my notes from that 1981 ICCB meeting online for History's sake.? The last page starts my notes for the Internet Meeting of 9/22, but JonP's minutes are much better. Here's a link for anyone who's interested: https://drive.google.com/file/d/1sxKhB-11jkVaVNCPVGwit3PTQDdta-Qr/view?usp=sharing /Jack Haverty On 6/2/26 08:53, D Waitzman via Internet-history wrote: > Regarding the history subtopic: > > https://www.iab.org/about/history/ > has: >> The origin of today?s IAB lies in the Internet Configuration Control Board (ICCB), which was created in 1979 by Vint Cerf, at that time program manager at DARPA, to advise him on technical issues. The ICCB was chaired by David Clark, MIT. >> >> In September 1984, after the ICCB meeting held at RSRE in Malvern, UK, the ICCB was disbanded and replaced by the Internet Advisory Board (IAB). This change was initiated by Dave Clark and Barry Leiner, who had taken over management of the Internet research program at DARPA. The IAB consisted of the chairs of the newly-formed research task forces and Jon Postel (ISI), as RFC editor and ?protocol czar?. The first set of chairs of the task forces were the members of the ICCB. The IAB was chaired by Dave Clark. > > I thought it was called the ICB, not ICCB. I attended a meeting in The Hague at NATO's STC around 1991. MAJ (or COL by then?) Pullen was funding us. > > Noel Chiappa wrote: >> There's a reason that, for professional historians, contemporaneous written records are the 'gold standard'. All other sources - especially human memory - are very inferior. > > Circa 1989 on a Friday afternoon, Steve Storch asked me to fly to California for a Monday ICB or IETF Host Requirements meeting to take notes. > Fortunately, BBN had a capable onsite travel agent in our building (10 Moulton) who quickly booked me tickets (pre-web this was a big deal). > I emailed out the next on Tuesday to Bob Braden, who emailed them out to the team with a subject line like "Shazam!" since the notes usually took longer to be written up and distributed. > (I was amused by the cost of a note taker for that contract.) > > Those certainly meet Noel's gold standard, but I cannot find that email. > > -- > David Waitzman > -------------- next part -------------- A non-text attachment was scrubbed... Name: OpenPGP_signature.asc Type: application/pgp-signature Size: 665 bytes Desc: OpenPGP digital signature URL: From craig at tereschau.net Tue Jun 2 12:24:34 2026 From: craig at tereschau.net (Craig Partridge) Date: Tue, 2 Jun 2026 13:24:34 -0600 Subject: [ih] capacity v bandwidth In-Reply-To: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> Message-ID: The last IETF that was entirely in one room was IETF 6 @ BBN (and I'm not sure it was all one room). It was a joint meeting with ANSI X3S3.3, which I remember as a series of plenary talks combined with very polite but pointed audience commentary about whether the Internet or OSI was leading the global networking charge. IETF 5 sought to create WGs, but with limited success. One afternoon we split into two working groups: one on BGP issues and one on network management. I turned out to be the only person in the network management meeting room:-). At some point Milo Medin took pity on me and took me out onto the Moffat Field runway to watch a U2 take off (still a treasured memory). Craig On Tue, Jun 2, 2026 at 11:28?AM Noel Chiappa via Internet-history < internet-history at elists.isoc.org> wrote: > > From: Barbara Denny > > > I doubt it was discussed in a hallway. To me the early meetings were > > more like a DARPA meeting ... Everyone was in one room as far as I > knew. > > The number of people attending was not that high. > > IETF meetings very quickly became fairly sizeable. From the meeting list I > linked to earlier: > > 1st IETF 21 Attendees January, 1986; San Diego > 2nd IETF 21 Attendees April, 1986; Aberdeen > 3rd IETF 18 Attendees July, 1986; Ann Arbor > 4th IETF 35 Attendees October, 1986; Menlo Park > 5th IETF 35 Attendees February, 1987; Moffett Field > 6th IETF 88 Attendees April, 1987; Boston > 7th IETF 101 Attendees July, 1987; McLean > 8th IETF 56 Attendees November, 1987; Boulder > 9th IETF 82 Attendees March, 1988; San Diego > 10th IETF 112 Attendees June, 1988; Annapolis > 11th IETF 114 Attendees October, 1988; Ann Arbor > 12th IETF 120 Attendees January, 1989; Austin > 13th IETF 114 Attendees April, 1989; Cocoa Beach > > With the timeframe that Brian deduced for the gateway/router change - > between > June 1988 and June 1989 - the initial growth had pretty clearly taken off. > > (The 7th, in McLean, was almost certainly the one where I remember Phill, > Dan > Lynch and I sitting in a bar at the end of the day chanting 'It's time to > get > real'; no doubt it being the first one with more than 100 attendees > contributedto that thinking.) > > Noel > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- ***** Craig Partridge's email account for professional society activities and mailing lists. From steve at shinkuro.com Tue Jun 2 13:43:30 2026 From: steve at shinkuro.com (Steve Crocker) Date: Tue, 2 Jun 2026 16:43:30 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> Message-ID: I've been enjoying these remembrances of the IETF's growth. Some years earlier, before the IETF was created, the Arpanet protocols were developed in the loosely organized Network Working Group (NWG). The early NWG meetings, Aug 1968 through early to mid 1969, started small and gradually grew. 6-8 people at first and then 10 to 12, IIRC. As the Arpanet came into existence and more nodes were planned or committed, more people showed up. Also, once the Host-Host protocol (later renamed Network Control Protocol) was defined, attention shifted to the "higher level" protocols: Telnet and FTP. I recall the moment I realized we needed two parallel meetings. I believe it happened during the SJCC conference in Atlantic City in 1970. I didn't keep records -- perhaps someone else did -- but the numbers were huge ;). There were surely more than 20 people, perhaps closer to 50. We clearly needed some organization and structure... Cheers, Steve On Tue, Jun 2, 2026 at 3:24?PM Craig Partridge via Internet-history < internet-history at elists.isoc.org> wrote: > The last IETF that was entirely in one room was IETF 6 @ BBN (and I'm not > sure it was all one room). It was a joint meeting with ANSI X3S3.3, which > I remember as a series of plenary talks combined with very polite but > pointed audience commentary about whether the Internet or OSI was leading > the global networking charge. > > IETF 5 sought to create WGs, but with limited success. One afternoon we > split into two working groups: one on BGP issues and one on network > management. I turned out to be the only person in the network management > meeting room:-). At some point Milo Medin took pity on me and took me out > onto the Moffat Field runway to watch a U2 take off (still a treasured > memory). > > Craig > > On Tue, Jun 2, 2026 at 11:28?AM Noel Chiappa via Internet-history < > internet-history at elists.isoc.org> wrote: > > > > From: Barbara Denny > > > > > I doubt it was discussed in a hallway. To me the early meetings > were > > > more like a DARPA meeting ... Everyone was in one room as far as I > > knew. > > > The number of people attending was not that high. > > > > IETF meetings very quickly became fairly sizeable. From the meeting list > I > > linked to earlier: > > > > 1st IETF 21 Attendees January, 1986; San Diego > > 2nd IETF 21 Attendees April, 1986; Aberdeen > > 3rd IETF 18 Attendees July, 1986; Ann Arbor > > 4th IETF 35 Attendees October, 1986; Menlo Park > > 5th IETF 35 Attendees February, 1987; Moffett Field > > 6th IETF 88 Attendees April, 1987; Boston > > 7th IETF 101 Attendees July, 1987; McLean > > 8th IETF 56 Attendees November, 1987; Boulder > > 9th IETF 82 Attendees March, 1988; San Diego > > 10th IETF 112 Attendees June, 1988; Annapolis > > 11th IETF 114 Attendees October, 1988; Ann Arbor > > 12th IETF 120 Attendees January, 1989; Austin > > 13th IETF 114 Attendees April, 1989; Cocoa Beach > > > > With the timeframe that Brian deduced for the gateway/router change - > > between > > June 1988 and June 1989 - the initial growth had pretty clearly taken > off. > > > > (The 7th, in McLean, was almost certainly the one where I remember Phill, > > Dan > > Lynch and I sitting in a bar at the end of the day chanting 'It's time to > > get > > real'; no doubt it being the first one with more than 100 attendees > > contributedto that thinking.) > > > > Noel > > -- > > Internet-history mailing list > > Internet-history at elists.isoc.org > > https://elists.isoc.org/mailman/listinfo/internet-history > > - > > Unsubscribe: > > > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > > > > -- > ***** > Craig Partridge's email account for professional society activities and > mailing lists. > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- Sent by a Verified sender From brian.e.carpenter at gmail.com Tue Jun 2 14:10:16 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Wed, 3 Jun 2026 09:10:16 +1200 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> Message-ID: <86b9bd11-c50f-48ba-a99e-6ed5c4551582@gmail.com> > before the IETF was created, the Arpanet protocols were developed > in the loosely organized Network Working Group (NWG) Which still survives as a ghost in some quite recent Internet-Drafts that still use an old (nroff?) template. See the first line in https://datatracker.ietf.org/doc/html/draft-klensin-std-numbers-03 Regards/Ng? mihi Brian Carpenter On 03-Jun-26 08:43, Steve Crocker via Internet-history wrote: > I've been enjoying these remembrances of the IETF's growth. Some years > earlier, before the IETF was created, the Arpanet protocols were developed > in the loosely organized Network Working Group (NWG). The early NWG > meetings, Aug 1968 through early to mid 1969, started small and gradually > grew. 6-8 people at first and then 10 to 12, IIRC. As the Arpanet > came into existence and more nodes were planned or committed, more people > showed up. Also, once the Host-Host protocol (later renamed Network > Control Protocol) was defined, attention shifted to the "higher level" > protocols: Telnet and FTP. I recall the moment I realized we needed two > parallel meetings. I believe it happened during the SJCC conference in > Atlantic City in 1970. I didn't keep records -- perhaps someone else did > -- but the numbers were huge ;). There were surely more than 20 people, > perhaps closer to 50. We clearly needed some organization and structure... > > Cheers, > > Steve > > > On Tue, Jun 2, 2026 at 3:24?PM Craig Partridge via Internet-history < > internet-history at elists.isoc.org> wrote: > >> The last IETF that was entirely in one room was IETF 6 @ BBN (and I'm not >> sure it was all one room). It was a joint meeting with ANSI X3S3.3, which >> I remember as a series of plenary talks combined with very polite but >> pointed audience commentary about whether the Internet or OSI was leading >> the global networking charge. >> >> IETF 5 sought to create WGs, but with limited success. One afternoon we >> split into two working groups: one on BGP issues and one on network >> management. I turned out to be the only person in the network management >> meeting room:-). At some point Milo Medin took pity on me and took me out >> onto the Moffat Field runway to watch a U2 take off (still a treasured >> memory). >> >> Craig >> >> On Tue, Jun 2, 2026 at 11:28?AM Noel Chiappa via Internet-history < >> internet-history at elists.isoc.org> wrote: >> >>> > From: Barbara Denny >>> >>> > I doubt it was discussed in a hallway. To me the early meetings >> were >>> > more like a DARPA meeting ... Everyone was in one room as far as I >>> knew. >>> > The number of people attending was not that high. >>> >>> IETF meetings very quickly became fairly sizeable. From the meeting list >> I >>> linked to earlier: >>> >>> 1st IETF 21 Attendees January, 1986; San Diego >>> 2nd IETF 21 Attendees April, 1986; Aberdeen >>> 3rd IETF 18 Attendees July, 1986; Ann Arbor >>> 4th IETF 35 Attendees October, 1986; Menlo Park >>> 5th IETF 35 Attendees February, 1987; Moffett Field >>> 6th IETF 88 Attendees April, 1987; Boston >>> 7th IETF 101 Attendees July, 1987; McLean >>> 8th IETF 56 Attendees November, 1987; Boulder >>> 9th IETF 82 Attendees March, 1988; San Diego >>> 10th IETF 112 Attendees June, 1988; Annapolis >>> 11th IETF 114 Attendees October, 1988; Ann Arbor >>> 12th IETF 120 Attendees January, 1989; Austin >>> 13th IETF 114 Attendees April, 1989; Cocoa Beach >>> >>> With the timeframe that Brian deduced for the gateway/router change - >>> between >>> June 1988 and June 1989 - the initial growth had pretty clearly taken >> off. >>> >>> (The 7th, in McLean, was almost certainly the one where I remember Phill, >>> Dan >>> Lynch and I sitting in a bar at the end of the day chanting 'It's time to >>> get >>> real'; no doubt it being the first one with more than 100 attendees >>> contributedto that thinking.) >>> >>> Noel >>> -- >>> Internet-history mailing list >>> Internet-history at elists.isoc.org >>> https://elists.isoc.org/mailman/listinfo/internet-history >>> - >>> Unsubscribe: >>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >>> >> >> >> -- >> ***** >> Craig Partridge's email account for professional society activities and >> mailing lists. >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > > From jack at 3kitty.org Tue Jun 2 15:43:17 2026 From: jack at 3kitty.org (Jack Haverty) Date: Tue, 2 Jun 2026 15:43:17 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> Message-ID: <06d4845f-d118-4c53-a1f2-5a58b1476e3b@3kitty.org> Yes, it's interesting to fill in gaps in the History.? I'm wondering about the evolution of the ICCB/IAB over the years.? I was on the ICCB from its inception until mid-1983, but don't know anything about its later history as the IAB. Through 1983, the ICCB was focussed on planning and prioritizing work towards the "January 1983" Internet, tasked to be one capable of handling heavy loads.??It wasn't just technical issues, although there were plenty of those.? Some had to be addressed before 1983 to create an operational Internet as a service, while others were classed as ongoing research topics since there was no consensus on any particular solution but lots of ideas to pursue. Most of the ICCB members were leading various Internet-related projects in their respective organizations, and would leave quarterly ICCB meetings with appropriate "marching orders" for their projects. But several of the ICCB members at that time were from other parts of the US DoD (not ARPA where Vint was), and likely dealt also with administrative or political tasks, such as getting TCP to become a DoD Standard, causing NBS/NIST certification procedures for TCP to be created so thaat government contractors and auditors could verify that they had implemented TCP, orchestrating the NCP/TCP cutover to occur on ARPANET, and undoubtedly many other non-technical activities that had to happen for the Internet to grow and expand and become a reliable communications service. My impression is that the role of the ICCB changed significantly over the years after it became the IAB, and its focus evolved to become almost purely technical.? But I don't know much at all about that part of the History. Anybody remember more? /Jack Haverty On 6/2/26 13:43, Steve Crocker via Internet-history wrote: > I've been enjoying these remembrances of the IETF's growth. Some years > earlier, before the IETF was created, the Arpanet protocols were developed > in the loosely organized Network Working Group (NWG). The early NWG > meetings, Aug 1968 through early to mid 1969, started small and gradually > grew. 6-8 people at first and then 10 to 12, IIRC. As the Arpanet > came into existence and more nodes were planned or committed, more people > showed up. Also, once the Host-Host protocol (later renamed Network > Control Protocol) was defined, attention shifted to the "higher level" > protocols: Telnet and FTP. I recall the moment I realized we needed two > parallel meetings. I believe it happened during the SJCC conference in > Atlantic City in 1970. I didn't keep records -- perhaps someone else did > -- but the numbers were huge ;). There were surely more than 20 people, > perhaps closer to 50. We clearly needed some organization and structure... > > Cheers, > > Steve > > > On Tue, Jun 2, 2026 at 3:24?PM Craig Partridge via Internet-history < > internet-history at elists.isoc.org> wrote: > >> The last IETF that was entirely in one room was IETF 6 @ BBN (and I'm not >> sure it was all one room). It was a joint meeting with ANSI X3S3.3, which >> I remember as a series of plenary talks combined with very polite but >> pointed audience commentary about whether the Internet or OSI was leading >> the global networking charge. >> >> IETF 5 sought to create WGs, but with limited success. One afternoon we >> split into two working groups: one on BGP issues and one on network >> management. I turned out to be the only person in the network management >> meeting room:-). At some point Milo Medin took pity on me and took me out >> onto the Moffat Field runway to watch a U2 take off (still a treasured >> memory). >> >> Craig >> >> On Tue, Jun 2, 2026 at 11:28?AM Noel Chiappa via Internet-history < >> internet-history at elists.isoc.org> wrote: >> >>> > From: Barbara Denny >>> >>> > I doubt it was discussed in a hallway. To me the early meetings >> were >>> > more like a DARPA meeting ... Everyone was in one room as far as I >>> knew. >>> > The number of people attending was not that high. >>> >>> IETF meetings very quickly became fairly sizeable. From the meeting list >> I >>> linked to earlier: >>> >>> 1st IETF 21 Attendees January, 1986; San Diego >>> 2nd IETF 21 Attendees April, 1986; Aberdeen >>> 3rd IETF 18 Attendees July, 1986; Ann Arbor >>> 4th IETF 35 Attendees October, 1986; Menlo Park >>> 5th IETF 35 Attendees February, 1987; Moffett Field >>> 6th IETF 88 Attendees April, 1987; Boston >>> 7th IETF 101 Attendees July, 1987; McLean >>> 8th IETF 56 Attendees November, 1987; Boulder >>> 9th IETF 82 Attendees March, 1988; San Diego >>> 10th IETF 112 Attendees June, 1988; Annapolis >>> 11th IETF 114 Attendees October, 1988; Ann Arbor >>> 12th IETF 120 Attendees January, 1989; Austin >>> 13th IETF 114 Attendees April, 1989; Cocoa Beach >>> >>> With the timeframe that Brian deduced for the gateway/router change - >>> between >>> June 1988 and June 1989 - the initial growth had pretty clearly taken >> off. >>> (The 7th, in McLean, was almost certainly the one where I remember Phill, >>> Dan >>> Lynch and I sitting in a bar at the end of the day chanting 'It's time to >>> get >>> real'; no doubt it being the first one with more than 100 attendees >>> contributedto that thinking.) >>> >>> Noel >>> -- >>> Internet-history mailing list >>> Internet-history at elists.isoc.org >>> https://elists.isoc.org/mailman/listinfo/internet-history >>> - >>> Unsubscribe: >>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> >> -- >> ***** >> Craig Partridge's email account for professional society activities and >> mailing lists. >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > -------------- next part -------------- A non-text attachment was scrubbed... Name: OpenPGP_signature.asc Type: application/pgp-signature Size: 665 bytes Desc: OpenPGP digital signature URL: From vgcerf at gmail.com Tue Jun 2 17:05:59 2026 From: vgcerf at gmail.com (vinton cerf) Date: Tue, 2 Jun 2026 20:05:59 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: References: Message-ID: There was also an ICB, as I recall chaired by Peter Kirstein and set up by Bob Kahn. ICB was International Cooperation (or Coordination?) Board - UK, Germany, Italy, US (possibly Norway?) were represented, I think. v On Tue, Jun 2, 2026 at 11:53?AM D Waitzman via Internet-history < internet-history at elists.isoc.org> wrote: > Regarding the history subtopic: > > https://www.iab.org/about/history/ > has: > > The origin of today?s IAB lies in the Internet Configuration Control > Board (ICCB), which was created in 1979 by Vint Cerf, at that time program > manager at DARPA, to advise him on technical issues. The ICCB was chaired > by David Clark, MIT. > > > > In September 1984, after the ICCB meeting held at RSRE in Malvern, UK, > the ICCB was disbanded and replaced by the Internet Advisory Board (IAB). > This change was initiated by Dave Clark and Barry Leiner, who had taken > over management of the Internet research program at DARPA. The IAB > consisted of the chairs of the newly-formed research task forces and Jon > Postel (ISI), as RFC editor and ?protocol czar?. The first set of chairs of > the task forces were the members of the ICCB. The IAB was chaired by Dave > Clark. > > > I thought it was called the ICB, not ICCB. I attended a meeting in The > Hague at NATO's STC around 1991. MAJ (or COL by then?) Pullen was funding > us. > > Noel Chiappa wrote: > > There's a reason that, for professional historians, contemporaneous > written records are the 'gold standard'. All other sources - especially > human memory - are very inferior. > > > Circa 1989 on a Friday afternoon, Steve Storch asked me to fly to > California for a Monday ICB or IETF Host Requirements meeting to take notes. > Fortunately, BBN had a capable onsite travel agent in our building (10 > Moulton) who quickly booked me tickets (pre-web this was a big deal). > I emailed out the next on Tuesday to Bob Braden, who emailed them out to > the team with a subject line like "Shazam!" since the notes usually took > longer to be written up and distributed. > (I was amused by the cost of a note taker for that contract.) > > Those certainly meet Noel's gold standard, but I cannot find that email. > > -- > David Waitzman > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From vgcerf at gmail.com Tue Jun 2 17:06:45 2026 From: vgcerf at gmail.com (vinton cerf) Date: Tue, 2 Jun 2026 20:06:45 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> Message-ID: there were only about 20 people at the San Diego 1986 meeting. v On Tue, Jun 2, 2026 at 1:28?PM Noel Chiappa via Internet-history < internet-history at elists.isoc.org> wrote: > > From: Barbara Denny > > > I doubt it was discussed in a hallway. To me the early meetings were > > more like a DARPA meeting ... Everyone was in one room as far as I > knew. > > The number of people attending was not that high. > > IETF meetings very quickly became fairly sizeable. From the meeting list I > linked to earlier: > > 1st IETF 21 Attendees January, 1986; San Diego > 2nd IETF 21 Attendees April, 1986; Aberdeen > 3rd IETF 18 Attendees July, 1986; Ann Arbor > 4th IETF 35 Attendees October, 1986; Menlo Park > 5th IETF 35 Attendees February, 1987; Moffett Field > 6th IETF 88 Attendees April, 1987; Boston > 7th IETF 101 Attendees July, 1987; McLean > 8th IETF 56 Attendees November, 1987; Boulder > 9th IETF 82 Attendees March, 1988; San Diego > 10th IETF 112 Attendees June, 1988; Annapolis > 11th IETF 114 Attendees October, 1988; Ann Arbor > 12th IETF 120 Attendees January, 1989; Austin > 13th IETF 114 Attendees April, 1989; Cocoa Beach > > With the timeframe that Brian deduced for the gateway/router change - > between > June 1988 and June 1989 - the initial growth had pretty clearly taken off. > > (The 7th, in McLean, was almost certainly the one where I remember Phill, > Dan > Lynch and I sitting in a bar at the end of the day chanting 'It's time to > get > real'; no doubt it being the first one with more than 100 attendees > contributedto that thinking.) > > Noel > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From vint at google.com Tue Jun 2 17:59:08 2026 From: vint at google.com (Vint Cerf) Date: Tue, 2 Jun 2026 20:59:08 -0400 Subject: [ih] capacity v bandwidth In-Reply-To: <4caa86b9-253e-48dd-bae6-863952a7c0e0@3kitty.org> References: <4caa86b9-253e-48dd-bae6-863952a7c0e0@3kitty.org> Message-ID: inline responses On Tue, Jun 2, 2026 at 1:39?PM Jack Haverty via Internet-history < internet-history at elists.isoc.org> wrote: > Not quite what I remember. > > The ICCB was the Internet Configuration Control Board. Bob Kahn called > it "Vint's Cabinet". The name was chosen explicitly to make it sound > bureaucratic and boring. Otherwise everyone would have wanted to be on > it, and a big problem at the time (late 1970s) was the size of the > various "internet meetings". Dave Clark was a member of the ICCB, but > Vint was clearly the Chair. I don't recall a single meeting not run by > Vint. > but officially Dave was the chair even if I ran the meeting. I had always intended to hand off to him, as I finally did in late 1982. Dave (among many others) was an extraordinarily clear thinker about the architecture and protocols of the Internet. Not better proof of that is his book "How to Design an Internet." > > The ICB was the International Collaboration Board (or something like > that). There was overlap (e.g., Vint) in the ICB and ICCB memberships. > The ICCB focussed on planning and execution of the steps needed to > target a "January 1983" Internet capable of handling "heavy load". The > ICB focussed on issues specific to trans-national connectivity, e.g., > between the US and Europe, and had members such as John Laws (RSRE), > Horst Klaussen (DFVLR), Peter Kirstein (UCL), and Paal Spilling (NDRE). > May have been others too. Someone from CNUCE (Italy) was probably > involved but I never met him or her. > Yes that's correct about ICB. Bob Kahn sert that up and I believe Peter K chaired. > > The earliest ICCB meeting I can find so far in my notes was on September > 21, 1981. I think it may have been the first meeting of the ICCB, based > on what's in the notes I took. I recall at some point around then > getting an email from Vint asking if I could come a day early to the > next Internet Meeting, which turned out to be the first meeting of the > new ICCB. I'm not sure, but think it was that September 1981 meeting. > My notes don't have an attendee list, but IIRC the ICCB meetings > included Jim Mathis, Dave Clark, Jon Postel, Dave Mills, Bob Braden, Ed > Cain, Ray McFarland, myself (Jack Haverty), and of course Vint. > > Meetings of the ICCB were usually held the day before a scheduled > Internet Meeting. We were typically considering what had to happen > near-term for a January 1983 target of an Internet capable of handling > "heavy load", as well as long-term issues that needed more research. > The 1983 target was clearly associated with the intent to replace NCP on > the ARPANET with TCP, and the adoption of TCP as a DoD Standard. > > It was less than two years away -- not a lot of time to get everything > in place. There was a clear conflict between "Research" and > "Operations". Research seeks to study and try new ideas; Operations > seeks to get something in place and working as a service. The Internet > needed to do both. > > After an ICCB meeting we all returned to our regular activities and did > whatever we could to get the near-term actions to happen and the > long-term issues to get as much attention as possible. The > prioritization helped also for Vint to control the Internet Meeting > which immediately followed. "Herding Cats" comes to mind.... > > I was on the ICCB until Vint's departure from ARPA and Barry Leiner > reorganized the ICCB to become the IAB. There had been a major > re-organization at BBN in July 1983 and I was no longer to be involved > in the "research" contracts. > > I can't remember if this list accepts attachments -- so I've put my > notes from that 1981 ICCB meeting online for History's sake. The last > page starts my notes for the Internet Meeting of 9/22, but JonP's > minutes are much better. > > Here's a link for anyone who's interested: > > > https://drive.google.com/file/d/1sxKhB-11jkVaVNCPVGwit3PTQDdta-Qr/view?usp=sharing AMAZING!!! > > > /Jack Haverty > > On 6/2/26 08:53, D Waitzman via Internet-history wrote: > > Regarding the history subtopic: > > > > https://www.iab.org/about/history/ > > has: > >> The origin of today?s IAB lies in the Internet Configuration Control > Board (ICCB), which was created in 1979 by Vint Cerf, at that time program > manager at DARPA, to advise him on technical issues. The ICCB was chaired > by David Clark, MIT. > >> > >> In September 1984, after the ICCB meeting held at RSRE in Malvern, UK, > the ICCB was disbanded and replaced by the Internet Advisory Board (IAB). > This change was initiated by Dave Clark and Barry Leiner, who had taken > over management of the Internet research program at DARPA. The IAB > consisted of the chairs of the newly-formed research task forces and Jon > Postel (ISI), as RFC editor and ?protocol czar?. The first set of chairs of > the task forces were the members of the ICCB. The IAB was chaired by Dave > Clark. > > > > I thought it was called the ICB, not ICCB. I attended a meeting in The > Hague at NATO's STC around 1991. MAJ (or COL by then?) Pullen was funding > us. > > > > Noel Chiappa wrote: > >> There's a reason that, for professional historians, contemporaneous > written records are the 'gold standard'. All other sources - especially > human memory - are very inferior. > > > > Circa 1989 on a Friday afternoon, Steve Storch asked me to fly to > California for a Monday ICB or IETF Host Requirements meeting to take notes. > > Fortunately, BBN had a capable onsite travel agent in our building (10 > Moulton) who quickly booked me tickets (pre-web this was a big deal). > > I emailed out the next on Tuesday to Bob Braden, who emailed them out to > the team with a subject line like "Shazam!" since the notes usually took > longer to be written up and distributed. > > (I was amused by the cost of a note taker for that contract.) > > > > Those certainly meet Noel's gold standard, but I cannot find that email. > > > > -- > > David Waitzman > > > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- After July 1, Please send any postal/overnight deliveries to: Vint Cerf 1435 Woodhurst Blvd McLean, VA 22102 Until July 1, send postal/overnight deliveries to: Google, LLC 1900 Reston Metro Plaza, 16th Floor Reston, VA 20190 +1 650-224-2788 From brian.e.carpenter at gmail.com Tue Jun 2 18:14:08 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Wed, 3 Jun 2026 13:14:08 +1200 Subject: [ih] capacity v bandwidth In-Reply-To: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> Message-ID: <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> The proceedings of those early meetings are on line as scanned PDF, at https://www.ietf.org/meeting/past/ For example, in the proceedings of IETF 9 (San Diego, March 1988) there are ~18 mentions of "router" and ~63 of "gateway". (Approximate numbers because there might be OCR errors in the PDF.) By IETF 17 (Pittsburgh, May 1990) there were ~94 "routers" and ~65 "gateways". Somebody asked about very old IETF mail archives. As far as I can tell, the IETF has an archive of its main list from Thu, 09 April 1992 01:51 UTC onwards, when the list was still at ietf at isi.edu. Any earlier archive would be of interest and I'm sure they would tack it on. https://mailarchive.ietf.org/arch/browse/ietf/?so=date (Thence, I was led to my own first-ever message to the IETF list: https://mailarchive.ietf.org/arch/msg/ietf/8XWGMC_iHsUWlbsnswyeuo6n_-8/ ) As of right now, that archive contains 146146 messages. Regards/Ng? mihi Brian Carpenter On 03-Jun-26 05:28, Noel Chiappa via Internet-history wrote: > > From: Barbara Denny > > > I doubt it was discussed in a hallway. To me the early meetings were > > more like a DARPA meeting ... Everyone was in one room as far as I knew. > > The number of people attending was not that high. > > IETF meetings very quickly became fairly sizeable. From the meeting list I > linked to earlier: > > 1st IETF 21 Attendees January, 1986; San Diego > 2nd IETF 21 Attendees April, 1986; Aberdeen > 3rd IETF 18 Attendees July, 1986; Ann Arbor > 4th IETF 35 Attendees October, 1986; Menlo Park > 5th IETF 35 Attendees February, 1987; Moffett Field > 6th IETF 88 Attendees April, 1987; Boston > 7th IETF 101 Attendees July, 1987; McLean > 8th IETF 56 Attendees November, 1987; Boulder > 9th IETF 82 Attendees March, 1988; San Diego > 10th IETF 112 Attendees June, 1988; Annapolis > 11th IETF 114 Attendees October, 1988; Ann Arbor > 12th IETF 120 Attendees January, 1989; Austin > 13th IETF 114 Attendees April, 1989; Cocoa Beach > > With the timeframe that Brian deduced for the gateway/router change - between > June 1988 and June 1989 - the initial growth had pretty clearly taken off. > > (The 7th, in McLean, was almost certainly the one where I remember Phill, Dan > Lynch and I sitting in a bar at the end of the day chanting 'It's time to get > real'; no doubt it being the first one with more than 100 attendees > contributedto that thinking.) > > Noel From gregskinner0 at icloud.com Wed Jun 3 10:11:28 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Wed, 3 Jun 2026 10:11:28 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> Message-ID: <08CE3358-45CB-42B1-9DF7-9A3C5136E165@icloud.com> On Jun 2, 2026, at 6:14?PM, Brian E Carpenter via Internet-history wrote: > > The proceedings of those early meetings are on line as scanned PDF, > at https://www.ietf.org/meeting/past/ > > For example, in the proceedings of IETF 9 (San Diego, March 1988) > there are ~18 mentions of "router" and ~63 of "gateway". > (Approximate numbers because there might be OCR errors in the PDF.) > > By IETF 17 (Pittsburgh, May 1990) there were ~94 "routers" and > ~65 "gateways". > > Somebody asked about very old IETF mail archives. As far as I can tell, > the IETF has an archive of its main list from Thu, 09 April 1992 01:51 UTC > onwards, when the list was still at ietf at isi.edu. Any earlier archive > would be of interest and I'm sure they would tack it on. > > https://mailarchive.ietf.org/arch/browse/ietf/?so=date > > (Thence, I was led to my own first-ever message to the IETF list: > https://mailarchive.ietf.org/arch/msg/ietf/8XWGMC_iHsUWlbsnswyeuo6n_-8/ ) > > As of right now, that archive contains 146146 messages. > > Regards/Ng? mihi > Brian Carpenter > There is some IETF list mail from December 1989 through December 1990 in the 1990-all file that is part of the collection that was formerly accessible via ftp. I don?t know of an older online repository for the IETF list. I also find that regrettable. The oldest mention of ?router? I was able to find in tcp-ip list mail that was gatewayed (lol) to USENET news was posted to mod.protocols.tcp-ip by Ron Natalie in 1985 . --gregbo From b_a_denny at yahoo.com Wed Jun 3 10:26:01 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Wed, 3 Jun 2026 17:26:01 +0000 (UTC) Subject: [ih] capacity v bandwidth In-Reply-To: <08CE3358-45CB-42B1-9DF7-9A3C5136E165@icloud.com> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> <08CE3358-45CB-42B1-9DF7-9A3C5136E165@icloud.com> Message-ID: <1108784415.793951.1780507561762@mail.yahoo.com> On Wednesday, June 3, 2026 at 10:11:58 AM PDT, Greg Skinner via Internet-history wrote: On Jun 2, 2026, at 6:14?PM, Brian E Carpenter via Internet-history wrote: > > The proceedings of those early meetings are on line as scanned PDF, > at https://www.ietf.org/meeting/past/ > > For example, in the proceedings of IETF 9 (San Diego, March 1988) > there are ~18 mentions of "router" and ~63 of "gateway". > (Approximate numbers because there might be OCR errors in the PDF.) > > By IETF 17 (Pittsburgh, May 1990) there were ~94 "routers" and > ~65 "gateways". > > Somebody asked about very old IETF mail archives. As far as I can tell, > the IETF has an archive of its main list from Thu, 09 April 1992 01:51 UTC > onwards, when the list was still at ietf at isi.edu. Any earlier archive > would be of interest and I'm sure they would tack it on. > > https://mailarchive.ietf.org/arch/browse/ietf/?so=date > > (Thence, I was led to my own first-ever message to the IETF list: > https://mailarchive.ietf.org/arch/msg/ietf/8XWGMC_iHsUWlbsnswyeuo6n_-8/ ) > > As of right now, that archive contains 146146 messages. > > Regards/Ng? mihi >? Brian Carpenter > There is some IETF list mail from December 1989 through December 1990 in the 1990-all file that is part of the collection that was formerly accessible via ftp.? I don?t know of an older online repository for the IETF list. I also find that regrettable. The oldest mention of ?router? I was able to find in tcp-ip list mail that was gatewayed (lol) to USENET news was posted to mod.protocols.tcp-ip by Ron Natalie in 1985 . --gregbo -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From b_a_denny at yahoo.com Wed Jun 3 11:47:38 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Wed, 3 Jun 2026 18:47:38 +0000 (UTC) Subject: [ih] capacity v bandwidth In-Reply-To: <1108784415.793951.1780507561762@mail.yahoo.com> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> <08CE3358-45CB-42B1-9DF7-9A3C5136E165@icloud.com> <1108784415.793951.1780507561762@mail.yahoo.com> Message-ID: <1696356041.835845.1780512458158@mail.yahoo.com> Oops.? Obviously I didn't mean to send the previous email out.? ?I was thinking about putting my thoughts down (Sometimes i use draft emails like a diary but never send them). Anyway, my preliminary thought is that the full context of? the word router in this email is for a service on the BRL gateway. The entity is still referred to as a gateway. I also used Gemini to check into the history of the? BRL vax. The story of the BRL vax it tells is interesting but probably has lots of problems (like linking it to packet? radio and logical addressing etc.? ?i haven't heard of any of this. I think it may be throwing stuff in based on my previous queries. I have been playing with it). BTW, is RIP still around? barbara On Wednesday, June 3, 2026 at 10:26:13 AM PDT, Barbara Denny via Internet-history wrote: ? ? On Wednesday, June 3, 2026 at 10:11:58 AM PDT, Greg Skinner via Internet-history wrote:? On Jun 2, 2026, at 6:14?PM, Brian E Carpenter via Internet-history wrote: > > The proceedings of those early meetings are on line as scanned PDF, > at https://www.ietf.org/meeting/past/ > > For example, in the proceedings of IETF 9 (San Diego, March 1988) > there are ~18 mentions of "router" and ~63 of "gateway". > (Approximate numbers because there might be OCR errors in the PDF > > By IETF 17 (Pittsburgh, May 1990) there were ~94 "routers" and > ~65 "gateways". > > Somebody asked about very old IETF mail archives. As far as I can tell, > the IETF has an archive of its main list from Thu, 09 April 1992 01:51 UTC > onwards, when the list was still at ietf at isi.edu. Any earlier archive > would be of interest and I'm sure they would tack it on. > > https://mailarchive.ietf.org/arch/browse/ietf/?so=date > > (Thence, I was led to my own first-ever message to the IETF list: > https://mailarchive.ietf.org/arch/msg/ietf/8XWGMC_iHsUWlbsnswyeuo6n_-8/ ) > > As of right now, that archive contains 146146 messages. > > Regards/Ng? mihi >? Brian Carpenter > There is some IETF list mail from December 1989 through December 1990 in the 1990-all file that is part of the collection that was formerly accessible via ftp.? I don?t know of an older online repository for the IETF list. I also find that regrettable. The oldest mention of ?router? I was able to find in tcp-ip list mail that was gatewayed (lol) to USENET news was posted to mod.protocols.tcp-ip by Ron Natalie in 1985 . --gregbo -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history ? -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From b_a_denny at yahoo.com Wed Jun 3 12:40:48 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Wed, 3 Jun 2026 19:40:48 +0000 (UTC) Subject: [ih] Possible new source of information References: <2134852620.851583.1780515648372.ref@mail.yahoo.com> Message-ID: <2134852620.851583.1780515648372@mail.yahoo.com> While poking around i found this NASA website. https://ntrs.nasa.gov/ I think SRI's funding sometimes came through NASA Ames so we would send contract deliverables to them as well as ARPA/DARPA (like the work on DARTnet).? I haven't had much luck using it yet but i haven't had time to really try it out. barbara From gregskinner0 at icloud.com Wed Jun 3 13:01:30 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Wed, 3 Jun 2026 13:01:30 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: <1696356041.835845.1780512458158@mail.yahoo.com> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> <08CE3358-45CB-42B1-9DF7-9A3C5136E165@icloud.com> <1108784415.793951.1780507561762@mail.yahoo.com> <1696356041.835845.1780512458158@mail.yahoo.com> Message-ID: <9C78C8E4-41DB-4B17-A994-B2430720F9CD@icloud.com> On Jun 3, 2026, at 11:47?AM, Barbara Denny via Internet-history wrote: > > Oops. Obviously I didn't mean to send the previous email out. I was thinking about putting my thoughts down (Sometimes i use draft emails like a diary but never send them). Anyway, my preliminary thought is that the full context of the word router in this email is for a service on the BRL gateway. The entity is still referred to as a gateway. I also used Gemini to check into the history of the BRL vax. The story of the BRL vax it tells is interesting but probably has lots of problems (like linking it to packet radio and logical addressing etc. i haven't heard of any of this. I think it may be throwing stuff in based on my previous queries. I have been playing with it). > BTW, is RIP still around? > barbara > On Wednesday, June 3, 2026 at 10:26:13 AM PDT, Barbara Denny via Internet-history wrote: > > > > On Wednesday, June 3, 2026 at 10:11:58 AM PDT, Greg Skinner via Internet-history wrote: > > On Jun 2, 2026, at 6:14?PM, Brian E Carpenter via Internet-history wrote: >> >> The proceedings of those early meetings are on line as scanned PDF, >> at https://www.ietf.org/meeting/past/ >> >> For example, in the proceedings of IETF 9 (San Diego, March 1988) >> there are ~18 mentions of "router" and ~63 of "gateway". >> (Approximate numbers because there might be OCR errors in the PDF >> >> By IETF 17 (Pittsburgh, May 1990) there were ~94 "routers" and >> ~65 "gateways". >> >> Somebody asked about very old IETF mail archives. As far as I can tell, >> the IETF has an archive of its main list from Thu, 09 April 1992 01:51 UTC >> onwards, when the list was still at ietf at isi.edu. Any earlier archive >> would be of interest and I'm sure they would tack it on. >> >> https://mailarchive.ietf.org/arch/browse/ietf/?so=date >> >> (Thence, I was led to my own first-ever message to the IETF list: >> https://mailarchive.ietf.org/arch/msg/ietf/8XWGMC_iHsUWlbsnswyeuo6n_-8/ ) >> >> As of right now, that archive contains 146146 messages. >> >> Regards/Ng? mihi >> Brian Carpenter >> > > There is some IETF list mail from December 1989 through December 1990 in the 1990-all file that is part of the collection that was formerly accessible via ftp. I don?t know of an older online repository for the IETF list. I also find that regrettable. > > The oldest mention of ?router? I was able to find in tcp-ip list mail that was gatewayed (lol) to USENET news was posted to mod.protocols.tcp-ip by Ron Natalie in 1985 . > > ?gregbo > I can?t say for sure what service Ron Natalie might have been referring to, but the use of ?router? in his tcp-ip post seems consistent with other uses of it at that time, to describe a user space or kernel function implemented on host machines that exchanged routing information. A couple of those ?routers? are actually described (with ?router? explicitly used) in RFC 898 (Paul Kirton?s and Mike Accetta?s). [1] For what it's worth, Mike Muuss? home page has some information about the BRL computing environment, including a picture of a VAX-11/780. [2] (There is a broken link to RFC 898.) Perhaps you or someone else (Noel or Bob Hinden, perhaps) who attended the RFC 898 meeting might remember if there was any particular mention of ?router? in this context. That might be something to look for in any ICCB or IAB mailing lists from that time period, if they were currently available. I don?t know of any current uses of RIP. I found a mention of it on the rtgwg asking for the removal of reference to it because it is almost dead nowadays (paraphrased). [3] --gregbo [1] https://www.rfc-editor.org/info/rfc898/ [2] https://web.archive.org/web/20000816192106/http://ftp.arl.army.mil/~mike/ [3] https://mailarchive.ietf.org/arch/msg/rtgwg/jWN5rIoqaY1CJgRTfA8KG0b5DS0/ From b_a_denny at yahoo.com Wed Jun 3 14:25:42 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Wed, 3 Jun 2026 21:25:42 +0000 (UTC) Subject: [ih] capacity v bandwidth In-Reply-To: <9C78C8E4-41DB-4B17-A994-B2430720F9CD@icloud.com> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> <08CE3358-45CB-42B1-9DF7-9A3C5136E165@icloud.com> <1108784415.793951.1780507561762@mail.yahoo.com> <1696356041.835845.1780512458158@mail.yahoo.com> <9C78C8E4-41DB-4B17-A994-B2430720F9CD@icloud.com> Message-ID: <495924448.901096.1780521942520@mail.yahoo.com> The only very, very vague memory i have of the router versus gateway topic is later. I think it came up when i was working? on the RP final report. I was talking to Jim Mathis.? It was more like what term should i use in the report. This is a very iffy memory so i don't think it is very useful or trustworthy. Jim left SRI in 1987. barbara On Wednesday, June 3, 2026 at 01:02:00 PM PDT, Greg Skinner wrote: On Jun 3, 2026, at 11:47?AM, Barbara Denny via Internet-history wrote: Oops.? Obviously I didn't mean to send the previous email out.? ?I was thinking about putting my thoughts down (Sometimes i use draft emails like a diary but never send them). Anyway, my preliminary thought is that the full context of? the word router in this email is for a service on the BRL gateway. The entity is still referred to as a gateway. I also used Gemini to check into the history of the? BRL vax. The story of the BRL vax it tells is interesting but probably has lots of problems (like linking it to packet? radio and logical addressing etc.? ?i haven't heard of any of this. I think it may be throwing stuff in based on my previous queries. I have been playing with it). BTW, is RIP still around? barbara ???On Wednesday, June 3, 2026 at 10:26:13 AM PDT, Barbara Denny via Internet-history wrote: ? ? ? On Wednesday, June 3, 2026 at 10:11:58 AM PDT, Greg Skinner via Internet-history wrote:? On Jun 2, 2026, at 6:14?PM, Brian E Carpenter via Internet-history wrote: The proceedings of those early meetings are on line as scanned PDF, at https://www.ietf.org/meeting/past/ For example, in the proceedings of IETF 9 (San Diego, March 1988) there are ~18 mentions of "router" and ~63 of "gateway". (Approximate numbers because there might be OCR errors in the PDF By IETF 17 (Pittsburgh, May 1990) there were ~94 "routers" and ~65 "gateways". Somebody asked about very old IETF mail archives. As far as I can tell, the IETF has an archive of its main list from Thu, 09 April 1992 01:51 UTC onwards, when the list was still at ietf at isi.edu. Any earlier archive would be of interest and I'm sure they would tack it on. https://mailarchive.ietf.org/arch/browse/ietf/?so=date (Thence, I was led to my own first-ever message to the IETF list: https://mailarchive.ietf.org/arch/msg/ietf/8XWGMC_iHsUWlbsnswyeuo6n_-8/ ) As of right now, that archive contains 146146 messages. Regards/Ng? mihi ? Brian Carpenter There is some IETF list mail from December 1989 through December 1990 in the 1990-all file that is part of the collection that was formerly accessible via ftp.? I don?t know of an older online repository for the IETF list. I also find that regrettable. The oldest mention of ?router? I was able to find in tcp-ip list mail that was gatewayed (lol) to USENET news was posted to mod.protocols.tcp-ip by Ron Natalie in 1985 . ?gregbo I can?t say for sure what service Ron Natalie might have been referring to, but the use of ?router? in his tcp-ip post seems consistent with other uses of it at that time, to describe a user space or kernel function implemented on host machines that exchanged routing information. ?A couple of those ?routers? are actually described (with ?router? explicitly used) in RFC 898 (Paul Kirton?s and Mike Accetta?s). [1] For what it's worth, Mike Muuss? home page has some information about the BRL computing environment, including a picture of a VAX-11/780. [2] (There is a broken link to RFC 898.) Perhaps you or someone else (Noel or Bob Hinden, perhaps) who attended the RFC 898 meeting might remember if there was any particular mention of ?router? in this context. ?That might be something to look for in any ICCB or IAB mailing lists from that time period, if they were currently available. I don?t know of any current uses of RIP. ?I found a mention of it on the rtgwg asking for the removal of reference to it because it is almost dead nowadays (paraphrased). [3] --gregbo [1]?https://www.rfc-editor.org/info/rfc898/[2]?https://web.archive.org/web/20000816192106/http://ftp.arl.army.mil/~mike/[3]?https://mailarchive.ietf.org/arch/msg/rtgwg/jWN5rIoqaY1CJgRTfA8KG0b5DS0/ From cdel at firsthand.net Fri Jun 5 03:27:45 2026 From: cdel at firsthand.net (Christian de Larrinaga) Date: Fri, 05 Jun 2026 11:27:45 +0100 Subject: [ih] capacity v bandwidth In-Reply-To: References: <4caa86b9-253e-48dd-bae6-863952a7c0e0@3kitty.org> Message-ID: <87y0gt1dum.fsf@firsthand.net> fascinating! Agenda top of page two's problem list "Architecture". So much for the no architecture meme? I am also interested to see Rob Cole's name under the UCL site. Such a great survival! C Vint Cerf via Internet-history writes: > inline responses > > On Tue, Jun 2, 2026 at 1:39?PM Jack Haverty via Internet-history < > internet-history at elists.isoc.org> wrote: > >> Not quite what I remember. >> >> The ICCB was the Internet Configuration Control Board. Bob Kahn called >> it "Vint's Cabinet". The name was chosen explicitly to make it sound >> bureaucratic and boring. Otherwise everyone would have wanted to be on >> it, and a big problem at the time (late 1970s) was the size of the >> various "internet meetings". Dave Clark was a member of the ICCB, but >> Vint was clearly the Chair. I don't recall a single meeting not run by >> Vint. >> > but officially Dave was the chair even if I ran the meeting. I had always > intended to hand off to him, as I finally did in late 1982. > Dave (among many others) was an extraordinarily clear thinker about the > architecture and protocols of the Internet. Not better proof of that is his > book "How to Design an Internet." > > >> >> The ICB was the International Collaboration Board (or something like >> that). There was overlap (e.g., Vint) in the ICB and ICCB memberships. >> The ICCB focussed on planning and execution of the steps needed to >> target a "January 1983" Internet capable of handling "heavy load". The >> ICB focussed on issues specific to trans-national connectivity, e.g., >> between the US and Europe, and had members such as John Laws (RSRE), >> Horst Klaussen (DFVLR), Peter Kirstein (UCL), and Paal Spilling (NDRE). >> May have been others too. Someone from CNUCE (Italy) was probably >> involved but I never met him or her. >> > Yes that's correct about ICB. Bob Kahn sert that up and I believe Peter K > chaired. > >> >> The earliest ICCB meeting I can find so far in my notes was on September >> 21, 1981. I think it may have been the first meeting of the ICCB, based >> on what's in the notes I took. I recall at some point around then >> getting an email from Vint asking if I could come a day early to the >> next Internet Meeting, which turned out to be the first meeting of the >> new ICCB. I'm not sure, but think it was that September 1981 meeting. >> My notes don't have an attendee list, but IIRC the ICCB meetings >> included Jim Mathis, Dave Clark, Jon Postel, Dave Mills, Bob Braden, Ed >> Cain, Ray McFarland, myself (Jack Haverty), and of course Vint. >> >> Meetings of the ICCB were usually held the day before a scheduled >> Internet Meeting. We were typically considering what had to happen >> near-term for a January 1983 target of an Internet capable of handling >> "heavy load", as well as long-term issues that needed more research. >> The 1983 target was clearly associated with the intent to replace NCP on >> the ARPANET with TCP, and the adoption of TCP as a DoD Standard. >> >> It was less than two years away -- not a lot of time to get everything >> in place. There was a clear conflict between "Research" and >> "Operations". Research seeks to study and try new ideas; Operations >> seeks to get something in place and working as a service. The Internet >> needed to do both. >> >> After an ICCB meeting we all returned to our regular activities and did >> whatever we could to get the near-term actions to happen and the >> long-term issues to get as much attention as possible. The >> prioritization helped also for Vint to control the Internet Meeting >> which immediately followed. "Herding Cats" comes to mind.... >> >> I was on the ICCB until Vint's departure from ARPA and Barry Leiner >> reorganized the ICCB to become the IAB. There had been a major >> re-organization at BBN in July 1983 and I was no longer to be involved >> in the "research" contracts. >> >> I can't remember if this list accepts attachments -- so I've put my >> notes from that 1981 ICCB meeting online for History's sake. The last >> page starts my notes for the Internet Meeting of 9/22, but JonP's >> minutes are much better. >> >> Here's a link for anyone who's interested: >> >> >> https://drive.google.com/file/d/1sxKhB-11jkVaVNCPVGwit3PTQDdta-Qr/view?usp=sharing > > > AMAZING!!! > >> >> >> /Jack Haverty >> >> On 6/2/26 08:53, D Waitzman via Internet-history wrote: >> > Regarding the history subtopic: >> > >> > https://www.iab.org/about/history/ >> > has: >> >> The origin of today?s IAB lies in the Internet Configuration Control >> Board (ICCB), which was created in 1979 by Vint Cerf, at that time program >> manager at DARPA, to advise him on technical issues. The ICCB was chaired >> by David Clark, MIT. >> >> >> >> In September 1984, after the ICCB meeting held at RSRE in Malvern, UK, >> the ICCB was disbanded and replaced by the Internet Advisory Board (IAB). >> This change was initiated by Dave Clark and Barry Leiner, who had taken >> over management of the Internet research program at DARPA. The IAB >> consisted of the chairs of the newly-formed research task forces and Jon >> Postel (ISI), as RFC editor and ?protocol czar?. The first set of chairs of >> the task forces were the members of the ICCB. The IAB was chaired by Dave >> Clark. >> > >> > I thought it was called the ICB, not ICCB. I attended a meeting in The >> Hague at NATO's STC around 1991. MAJ (or COL by then?) Pullen was funding >> us. >> > >> > Noel Chiappa wrote: >> >> There's a reason that, for professional historians, contemporaneous >> written records are the 'gold standard'. All other sources - especially >> human memory - are very inferior. >> > >> > Circa 1989 on a Friday afternoon, Steve Storch asked me to fly to >> California for a Monday ICB or IETF Host Requirements meeting to take notes. >> > Fortunately, BBN had a capable onsite travel agent in our building (10 >> Moulton) who quickly booked me tickets (pre-web this was a big deal). >> > I emailed out the next on Tuesday to Bob Braden, who emailed them out to >> the team with a subject line like "Shazam!" since the notes usually took >> longer to be written up and distributed. >> > (I was amused by the cost of a note taker for that contract.) >> > >> > Those certainly meet Noel's gold standard, but I cannot find that email. >> > >> > -- >> > David Waitzman >> > >> >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > > > -- > After July 1, Please send any postal/overnight deliveries to: > Vint Cerf > 1435 Woodhurst Blvd > McLean, VA 22102 > > Until July 1, send postal/overnight deliveries to: > Google, LLC > 1900 Reston Metro Plaza, 16th Floor > Reston, VA 20190 > +1 650-224-2788 -- Christian de Larrinaga From geoff at iconia.com Mon Jun 8 11:45:12 2026 From: geoff at iconia.com (the keyboard of geoff goodfellow) Date: Mon, 8 Jun 2026 11:45:12 -0700 Subject: [ih] There are now more bots than humans on the internet. (Elias Al) Message-ID: For the first time in history. Cloudflare just confirmed it. Bots and AI agents now generate more web traffic than humans for the first time in internet history. Cloudflare CEO Matthew Prince described it as a major turning point. Automated bot requests account for roughly 57% of traffic to ordinary webpages worldwide, compared with about 43% generated by humans. And the CEO who announced it did not do so with a polished press release or a prepared statement. He posted four words on X on June 3, 2026: "Welp, that happened faster." [ https://x.com/eastdakota/status/2062212701414187452] Here is the full context behind those four words. Matthew Prince had previously forecast the bot-human crossover would occur by the end of 2027. He revised that to early 2027. Then agentic AI traffic grew so fast that the milestone arrived 18 months ahead of schedule in June 2026 catching even the CEO of the company tracking it by surprise. Here is what drove this faster than anyone predicted. The main driver is agentic AI, autonomous programs that browse the web on behalf of assistants like ChatGPT and Gemini. Before the generative AI era, bot traffic sat at around 20% of all web activity, with Google's web crawler serving as the largest single source. It is now 57.5%. 20% to 57.5%. In under three years. Here is the number that makes this even more alarming. Cloudflare's 2026 Threat Intelligence Report found that bots now account for 94% of all login attempts across its network, meaning only 6% of login attempts come from actual humans trying to sign in. 94% of every login attempt on the web. Bots. 6% of every login attempt. Real people. The infrastructure that was built to verify human identity is now processing mostly machine traffic. Here is the nuance worth understanding before the panic sets in. While bots now dominate HTML request traffic reading pages, scraping content, indexing sites humans still account for roughly 65% of total web activity when the metric expands to include app usage, video streaming, maps, and social media scrolling. Bots have overtaken humans in the specific act of navigating and reading the web, but not in the broader measure of people actually using the internet. And here is the question nobody has answered yet but everyone is now asking. Prince previously asked what pays for the web when more of its users are bots. Now that bots have crossed the majority line, that question is no longer theoretical. The entire economic model of the internet was built on human attention. Human clicks. Human eyeballs reading ads, buying products, subscribing to services, and generating revenue for every website, publisher, and platform online. The advertising model depends on humans seeing ads. E-commerce depends on humans making purchases. Subscription models depend on humans finding value. Analytics depend on humans generating meaningful engagement signals. The shift matters to anyone who publishes online, pays for hosting, or relies on an AI assistant that quietly fetches pages on their behalf, the economic assumptions the web was built on, advertising, referral clicks, and human attention, are being rewritten in real time. Sites can keep giving machines free access. Block them and lose referral traffic. Or charge them and the infrastructure to charge them now exists. None of those options are simple. None of them have been chosen at scale. And the bots keep coming regardless. Bot traffic has held between 53% and 60% in the weeks since the crossover. Prince said the actual crossover occurred in the last few months, though the data is messy enough that pinning down an exact date is difficult. We are clearly on the other side now, he added. Elon Musk replied to Prince's post with one word. "Wow." The internet was built for humans. For the first time in its history most of it is not being used by them. Source: Cloudflare ? Matthew Prince ? Search Engine Land ? Tom's Hardware ? TechTimes ? June 3?5, 2026 https://x.com/iam_elias1/status/2063690208961314996 -- Geoff.Goodfellow at iconia.com living as The Truth is True From gregskinner0 at icloud.com Sun Jun 14 15:02:58 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Sun, 14 Jun 2026 15:02:58 -0700 Subject: [ih] capacity v bandwidth In-Reply-To: <495924448.901096.1780521942520@mail.yahoo.com> References: <20260602172828.3E3EB18C074@mercury.lcs.mit.edu> <9d9f2ddc-57a6-4a84-8af4-02d59860cf52@gmail.com> <08CE3358-45CB-42B1-9DF7-9A3C5136E165@icloud.com> <1108784415.793951.1780507561762@mail.yahoo.com> <1696356041.835845.1780512458158@mail.yahoo.com> <9C78C8E4-41DB-4B17-A994-B2430720F9CD@icloud.com> <495924448.901096.1780521942520@mail.yahoo.com> Message-ID: On Jun 3, 2026, at 2:25?PM, Barbara Denny via Internet-history wrote: > > The only very, very vague memory i have of the router versus gateway topic is later. I think it came up when i was working on the RP final report. I was talking to Jim Mathis. It was more like what term should i use in the report. This is a very iffy memory so i don't think it is very useful or trustworthy. Jim left SRI in 1987. > barbara > This topic has come up on this list before; Brian Carpenter noted the use of ?router? in the 1979 PARC PUP internetwork architecture paper. [1] Some early 1980s Unix routed code and configuration information indicates it is based on XNS, and also has some mentions of ?router?. [2] [3] [4] [1] https://www.bitsavers.org/pdf/xerox/parc/techReports/CSL-79-10_Pup_An_Internetwork_Architecture_Jul79.pdf [2] https://www.tuhs.org/cgi-bin/utree.pl?file=2.9BSD/usr/net/src/netser/routed/routed.c [3] https://www.tuhs.org/cgi-bin/utree.pl?file=2.9BSD/usr/net/src/netser/routed/router.h [4] https://www.tuhs.org/cgi-bin/utree.pl?file=4.1cBSD/etc/services From jack at 3kitty.org Mon Jun 22 12:50:37 2026 From: jack at 3kitty.org (Jack Haverty) Date: Mon, 22 Jun 2026 12:50:37 -0700 Subject: [ih] History of AI and Internet Message-ID: With AI now in the news all the time, I've been wondering about the history of AI in The Internet, i.e., how AI technology has been used within the Internet over the years, rather than just a service which people can access over The Internet.? ?Personally I only know of a few such uses of AI in a communications role, quite a while ago. ?But maybe others remember more...? In the mid-1970s, I was involved in an AI project at MIT in Licklider's group, sponsored by ARPA.? ?It wasn't really part of the ARPANET but it did relate to communications.? ?The project involved using AI techniques of that era, then called "expert systems", to have computers decode hand-sent Morse radio transmissions.? ?The Intelligence Community was apparently quite interested in this problem at that time, and the project ran for several years and was deemed a success. FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 for info on that AI project.? Personally, I think of that 1990 report as an enumeration of "ARPA's Greatest Hits".? ?There are six projects listed from the Information Processing Techniques Office of ARPA (IPTO, where all the networking and computer projects happened).? Another project of those six IPTO successes is The ARPANET, described in Chapter 20.? ?The Internet was probably still too immature to be included at the time . Similarly, while I was at BBN in the 1980s there were tools developed to visualize activity in the ARPANET, and do tasks involved in network design - figuring out where new lines were needed, reconfiguring the topology of the ARPANET to address changes in traffic patterns, and other such analyses.? I'm not sure those projects would be recognized as "AI" today, but they were widely used to manage a variety of networks such as the ARPANET and DDN. The idea was to use computers to augment people skills, as it is in today's AI as well. Bob Kahn and I had a discussion at some point in late 1982 about AI and networks in ARPA's world.? ?Bob was aware of the ARPA-sponsored project done at MIT in the mid 1970s which used AI "expert system" techniques to decode Morse code.? ?We mused that a similar expert system approach could be used to manage networks, especially the emerging Internet which was much more complex than the ARPANET and had few tools available for operations and management. The concept was to use the plentiful supply of BBN engineers who had been debugging Internet crises for several years as the "experts", observing how they worked, and translating what they did into software to do tasks "automatically".? ?That started the "Automated Network Management" (ANM) project as a new research task at BBN.? ?I wrote the proposal to ARPA and waited for the contract to be signed.? But when BBN reorganized in July 1983 the ANM contract and I went separate ways so I never got to work on ANM and I don't think there was enough detail in the proposal to convey the concept onwards to the new project team. It strikes me that the real driver of AI today has evolved only partly from advances in algorithms such as LLMs.? I think another major factor has been the massive drop in computing costs over the last half-century.? AI systems, such as that Morse Code project, were affordable only by well-funded parts of government in the 1970s, requiring millions of dollars/pounds/rubles/yen/lira to field a system that provided minimal capability.? Today, the system that cost millions in the 1970s now costs pennies by comparison. Anybody else know more about the history of the use of AI *within* (not just on top of) The Internet?? ?For example, when there are problems in today's Internet, are AI techniques and tools used to diagnose and repair them?? ?What's the History of such things? /Jack Haverty -------------- next part -------------- A non-text attachment was scrubbed... Name: OpenPGP_signature.asc Type: application/pgp-signature Size: 665 bytes Desc: OpenPGP digital signature URL: From byeager at fastmail.fm Mon Jun 22 14:12:03 2026 From: byeager at fastmail.fm (william yeager) Date: Mon, 22 Jun 2026 14:12:03 -0700 Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: <14D3E77E-EDDB-4F8C-BDDA-268C54807969@fastmail.fm> Hi Jack, The answer is yes. And one needs to ask what was/is AI to uncover the history. As you?ve clearly pointed out, AI has a long history. It evolved with multiple versions boot strapped in 1955 with John McCarthy?s vision in which he coined the name AI with the following statement. "We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.? AI was never a silo and strongly depended on the evolution of multiple technical and non-technical advances. We all are aware of them. To be brief and answer your question below. A couple of years ago I saw a demo given by a member of DataBricks.com, a Big Data Analytics Company, engineering team. It used the company's Apache Spark APIs to detect, isolate, and stop an attack in real time. Apache Spark : "Apache Spark is a multi-language engine for executing data engineering, data science, and machine learning on single-node machines or clusters.? I also know of a small, new startup that is in stealth mode and whose goal is to use a "Small Language Model? (SLM) to secure businesses from Internet attacks. Why ?Small?? Because they use LLMs technics but training data is limited to the knowledge necessary to accomplish their goal. They are exclusively trained by professionals in their domain. They do not ?hallucinate.? A typical SML example is detecting breast cancer from mammograms. Slightly off topic and I apologize. > On Jun 22, 2026, at 12:50 PM, Jack Haverty via Internet-history wrote: > > Anybody else know more about the history of the use of AI *within* (not just on top of) The Internet? For example, when there are problems in today's Internet, are AI techniques and tools used to diagnose and repair them? What's the History of such things? ?????????? ????? ???? ????????? From b_a_denny at yahoo.com Mon Jun 22 19:42:47 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Tue, 23 Jun 2026 02:42:47 +0000 (UTC) Subject: [ih] History of AI and Internet In-Reply-To: <14D3E77E-EDDB-4F8C-BDDA-268C54807969@fastmail.fm> References: <14D3E77E-EDDB-4F8C-BDDA-268C54807969@fastmail.fm> Message-ID: <1583157813.1120712.1782182567385@mail.yahoo.com> Off hand i don't know of any specific projects (much less products)? but considering your question i am thinking of a couple things that happened in the 90s that might have included AI in some way. If i wanted to see what might have been done? ? i would start checking out the work done with software radios ( and perhaps composable systems).? Then there is cognitive radio.? Beyond spectrum issues there may have been AI-like things people thought about trying and did try. The other topic is Active Networks.? It was a Darpa Program headed by Doug Maughan i think.? I don't know what was really done.? I left SRI before the program really got off the ground so i don't know what the projects really were about.?? Now what is actually used? in the commercial internet is a different question. No insight there. Of course these suggestions might not be enough internet related for you. barbara On Monday, June 22, 2026 at 02:12:13 PM PDT, william yeager via Internet-history wrote: Hi Jack, The answer is yes. And one needs to ask what was/is AI to uncover the history. As you?ve clearly pointed out, AI has a long history. It evolved with multiple versions boot strapped in 1955 with John McCarthy?s vision in which he coined the name AI with the following statement. "We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.? AI was never a silo and strongly depended on the evolution of multiple technical and non-technical advances. We all are aware of them. To be brief and answer your question below. A couple of years ago I saw a demo given by a member of DataBricks.com, a Big Data Analytics Company, engineering team. It used the company's Apache Spark APIs to detect, isolate, and stop an attack in real time. Apache Spark : "Apache Spark is a multi-language engine for executing data engineering, data science, and machine learning on single-node machines or clusters.? I also know of a small, new startup that is in stealth mode and whose goal is to use a "Small Language Model? (SLM) to secure businesses from Internet attacks. Why ?Small?? Because they use LLMs technics but training data is limited to the knowledge necessary to accomplish their goal. They are exclusively trained by professionals in their domain. They do not ?hallucinate.? A typical SML example is detecting breast cancer from mammograms. Slightly off topic and I apologize. > On Jun 22, 2026, at 12:50 PM, Jack Haverty via Internet-history wrote: > > Anybody else know more about the history of the use of AI *within* (not just on top of) The Internet?? For example, when there are problems in today's Internet, are AI techniques and tools used to diagnose and repair them?? What's the History of such things? ?????????? ????? ???? ????????? -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From brian.e.carpenter at gmail.com Mon Jun 22 19:54:06 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Tue, 23 Jun 2026 14:54:06 +1200 Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: > Anybody else know more about the history of the use of AI *within* (not > just on top of) The Internet? For example, when there are problems in > today's Internet, are AI techniques and tools used to diagnose and > repair them? What's the History of such things? 1) For anyone interested in the history of AI, as well as how AI actually works, I recommend Anil Ananthaswamy's "Why Machines Learn" [1]. It's quite hard going if you want to follow all the maths, but it recapitualates the history of AI very well. 2) In answer to Jack's question, I think very little AI has been applied to networking in the past, but this is changing *very* fast. When we designed the GRASP protocol for autonomic networking, it was aimed at "intelligent" agents but we didn't have any - and that was only 5 years ago. Now the IETF is littered with drafts about intelligent agents. Big changes are coming. I have no doubt that 'expert systems' techniques have been applied, especially to network design, but expert systems were not real AI; they did incorporate heuristics, and anything using heuristics could be called AI at one stage. Not any more. A year ago I gave a talk about AI for network management [2] (aimed at final year students). I haven't checked recently, but I assume it's way out of date. [1] https://www.penguin.co.uk/books/446849/why-machines-learn-by-ananthaswamy-anil/9781802060867 [2] https://cs.auckland.ac.nz/~brian/CS389-BC-AI-netwk2025.pdf Regards/Ng? mihi Brian Carpenter On 23-Jun-26 07:50, Jack Haverty via Internet-history wrote: > With AI now in the news all the time, I've been wondering about the > history of AI in The Internet, i.e., how AI technology has been used > within the Internet over the years, rather than just a service which > people can access over The Internet.? ?Personally I only know of a few > such uses of AI in a communications role, quite a while ago. ?But maybe > others remember more...? > > In the mid-1970s, I was involved in an AI project at MIT in Licklider's > group, sponsored by ARPA.? ?It wasn't really part of the ARPANET but it > did relate to communications.? ?The project involved using AI techniques > of that era, then called "expert systems", to have computers decode > hand-sent Morse radio transmissions.? ?The Intelligence Community was > apparently quite interested in this problem at that time, and the > project ran for several years and was deemed a success. > > FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 for > info on that AI project.? Personally, I think of that 1990 report as an > enumeration of "ARPA's Greatest Hits".? ?There are six projects listed > from the Information Processing Techniques Office of ARPA (IPTO, where > all the networking and computer projects happened).? Another project of > those six IPTO successes is The ARPANET, described in Chapter 20.? ?The > Internet was probably still too immature to be included at the time . > > Similarly, while I was at BBN in the 1980s there were tools developed to > visualize activity in the ARPANET, and do tasks involved in network > design - figuring out where new lines were needed, reconfiguring the > topology of the ARPANET to address changes in traffic patterns, and > other such analyses.? I'm not sure those projects would be recognized as > "AI" today, but they were widely used to manage a variety of networks > such as the ARPANET and DDN. The idea was to use computers to augment > people skills, as it is in today's AI as well. > > Bob Kahn and I had a discussion at some point in late 1982 about AI and > networks in ARPA's world.? ?Bob was aware of the ARPA-sponsored project > done at MIT in the mid 1970s which used AI "expert system" techniques to > decode Morse code.? ?We mused that a similar expert system approach > could be used to manage networks, especially the emerging Internet which > was much more complex than the ARPANET and had few tools available for > operations and management. > > The concept was to use the plentiful supply of BBN engineers who had > been debugging Internet crises for several years as the "experts", > observing how they worked, and translating what they did into software > to do tasks "automatically".? ?That started the "Automated Network > Management" (ANM) project as a new research task at BBN.? ?I wrote the > proposal to ARPA and waited for the contract to be signed.? But when BBN > reorganized in July 1983 the ANM contract and I went separate ways so I > never got to work on ANM and I don't think there was enough detail in > the proposal to convey the concept onwards to the new project team. > > It strikes me that the real driver of AI today has evolved only partly > from advances in algorithms such as LLMs.? I think another major factor > has been the massive drop in computing costs over the last > half-century.? AI systems, such as that Morse Code project, were > affordable only by well-funded parts of government in the 1970s, > requiring millions of dollars/pounds/rubles/yen/lira to field a system > that provided minimal capability.? Today, the system that cost millions > in the 1970s now costs pennies by comparison. > > Anybody else know more about the history of the use of AI *within* (not > just on top of) The Internet?? ?For example, when there are problems in > today's Internet, are AI techniques and tools used to diagnose and > repair them?? ?What's the History of such things? > > /Jack Haverty > > From byeager at fastmail.fm Mon Jun 22 22:25:01 2026 From: byeager at fastmail.fm (william yeager) Date: Mon, 22 Jun 2026 22:25:01 -0700 Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: > On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history wrote: > > but expert systems were not real AI Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. I worked in the Stanford Knowledge Systems Lab at Stanford as well as for the DENDRAL project - AI in chemistry. My area of interest was using organic chemistry along with signal processing to analyze GCMS data with peak analysis gradients to detect inherited rare diseases in children from the fractions of these children?s blood. It worked and these were not heuristics. The spectra that were produced were give to several Phd?s in chemistry to identify the compounds, they were stored in a library for future recognition. Is it AI now? No. Was it then. Yes. What we did was state-of-the-art AI between 1970 and 1990 or so. Ed Feigenbaum who is a good friend received the Turing Prize along with Raj Redy for their work on expert systems in 1994. I am currently writing a history of AI from 1955 to the present from the perspective that it was a Gestalt. Have looked into what was AI for each of the multiple versions over the years. I?ve been in computer science and math for 60 years, lived through these, epochs, and believe me there were battles noted in the press initiated by antagonists who were worried that expert systems were going to replace the experts. AI evolved over the years to where we are now and definitely NOT as a SILO. Bill ?????????? ????? ???? ????????? From brian.e.carpenter at gmail.com Mon Jun 22 23:10:11 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Tue, 23 Jun 2026 18:10:11 +1200 Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> On 23-Jun-26 17:25, william yeager wrote: > >> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history > wrote: >> >> but expert systems were not real AI > > Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. Fair enough, but all the expert systems I saw were basically rules created by human experts. > I worked in the Stanford Knowledge Systems Lab at Stanford as well as for the DENDRAL project - AI in chemistry. My area of interest was using organic chemistry along with signal processing to analyze GCMS data with peak analysis gradients to detect inherited rare diseases in children from the fractions of these children?s blood. It worked and these were not heuristics. But were they systems that learned, and that detected features that they hadn't been told about in advance? > The spectra that were produced were give to several Phd?s in chemistry to identify the compounds, they were stored in a library for future recognition. Is it AI now? No. Was it then. Yes. What we did was state-of-the-art AI between 1970 and 1990 or so. > > Ed Feigenbaum who is a good friend received the Turing Prize along with Raj Redy for their work on expert systems in 1994. Yes, and I didn't mean to insult all that work (so I should have chosen my words more carefully). > I am currently writing a history of AI from 1955 to the present from the perspective that it was a Gestalt. Have looked into what was AI for each of the multiple versions over the years. I?ve been in computer science and math for 60 years, lived through these, epochs, and believe me there were battles noted in the press initiated by antagonists who were worried that expert systems were going to replace the experts. I never thought that, but it's already happening to a considerable extent with LLMs. > AI evolved over the years to where we are now and definitely NOT as a SILO. However, multi-layer perceptrons were basically ignored until a few years ago. (That includes me ignoring them when working on speech recognition in the late 1960s, because they were computationally infeasible.) Peace Brian > > Bill > > ?????????? > ????? > > ???? ????????? > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > From jeanjour at comcast.net Tue Jun 23 06:52:06 2026 From: jeanjour at comcast.net (John Day) Date: Tue, 23 Jun 2026 09:52:06 -0400 Subject: [ih] History of AI and Internet In-Reply-To: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> Message-ID: <55D00418-C37F-4DCD-98C6-1B7870C41525@comcast.net> I was told once that there are two forms of AI: 1) Those that are trying to understand and then emulate how the brain works, and 2) Those that are just trying to create something that acts like the brain. (I guess pass the Turing Test, but we are way beyond that.) Are we talking about the second case as AI? Is there any real progress on the first? Take care, John > On Jun 23, 2026, at 02:10, Brian E Carpenter via Internet-history wrote: > > On 23-Jun-26 17:25, william yeager wrote: >>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history > wrote: >>> >>> but expert systems were not real AI >> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. > > Fair enough, but all the expert systems I saw were basically rules created by human experts. > >> I worked in the Stanford Knowledge Systems Lab at Stanford as well as for the DENDRAL project - AI in chemistry. My area of interest was using organic chemistry along with signal processing to analyze GCMS data with peak analysis gradients to detect inherited rare diseases in children from the fractions of these children?s blood. It worked and these were not heuristics. > > But were they systems that learned, and that detected features that they hadn't been told about in advance? > >> The spectra that were produced were give to several Phd?s in chemistry to identify the compounds, they were stored in a library for future recognition. Is it AI now? No. Was it then. Yes. What we did was state-of-the-art AI between 1970 and 1990 or so. >> Ed Feigenbaum who is a good friend received the Turing Prize along with Raj Redy for their work on expert systems in 1994. > > Yes, and I didn't mean to insult all that work (so I should have chosen my words more carefully). > >> I am currently writing a history of AI from 1955 to the present from the perspective that it was a Gestalt. Have looked into what was AI for each of the multiple versions over the years. I?ve been in computer science and math for 60 years, lived through these, epochs, and believe me there were battles noted in the press initiated by antagonists who were worried that expert systems were going to replace the experts. > > I never thought that, but it's already happening to a considerable extent with LLMs. > >> AI evolved over the years to where we are now and definitely NOT as a SILO. > > However, multi-layer perceptrons were basically ignored until a few years ago. (That includes me ignoring them when working on speech recognition in the late 1960s, because they were computationally infeasible.) > > Peace > Brian > >> Bill >> ?????????? >> ????? >> >> ???? ????????? >> > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From byeager at fastmail.fm Tue Jun 23 08:10:43 2026 From: byeager at fastmail.fm (william yeager) Date: Tue, 23 Jun 2026 08:10:43 -0700 Subject: [ih] History of AI and Internet In-Reply-To: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> Message-ID: <93F30632-2721-44B5-852D-8DA3369212BF@fastmail.fm> Don?t want to spend too much time on this but apparently you?ve never read John McCarthy?s proposal for naming AI and what AI should do. Here it is: We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. > On Jun 22, 2026, at 11:10 PM, Brian E Carpenter wrote: > > On 23-Jun-26 17:25, william yeager wrote: >>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history > wrote: >>> >>> but expert systems were not real AI >> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. > > Fair enough, but all the expert systems I saw were basically rules created by human experts. This is true and the Expert Systems did simulate human thinking. These were not simple rules. Ted Shortliffe?s MYCIN and EMYCIN incorporated rules for diagnosing and prescribing antibiotics. These systems did as well as 90% of physicians. I personally worked on T-Helper that was used to remotely diagnose and treat AIDs patients and could be used by rural communities that lacked the expertise. The ideas weren?t mine. I helped a Phd student in medical informatics - ie - the student was an MD and wanted to add a Phd in the computational aspect of medicine - the underlying code misused TCP/IP - the latter was buried in the bowels of the large system and as I dug into it I also began to underside the medical side. Amazing software. The work in Expert Systems covered so many bases and Stanford the a national resource for the work in experimentation in AI and Medicine in the 1970s-90s. Yes, Expert Systems did simulate human thinking. This was not about generating stochastic sentences like LLMs do. I really understand the internals of these systems. For me the math was easy since I did my Phd studies in math at U of Washington. I am amazed at the beauty in LLMs and with the arrival in 2017 of the Foundation that enabled deep learning to scale and compute on 20 thousand NVIDIA GPUs in parallel; multiplying token based, weighted matrices on the order of 4096x4096 or larger along with smaller queries, keys and values matrices all created during training, the the embedded matrix 80000x4096 tokens + a few more ? the details are a thing of pure mathematical wonder to me. Do note that if one understands and not just knows advanced multivariate calculus and linear algebra along with standard error calculations and a few other details, they ought to be able to figure out how LLMs work. WHY is more complicated and worth another long discussion. The basic thing to understand about LLMs is that they are 90% training and 10% execution. They have a fixed memory of about 30 billion parameters and are becoming larger. The do not keep the vast amount of text that is scrapped off of the Internet; They have a fixed vocabulary of about 60 thousand words; and about 80 thousand tokens - the embedded array - on which all computation is based. Tokens are like syllables - pieces of words - have binary values that map to the text. So Bill might be ?bi? and ?ll?. One knows a token by the friends it keeps. They attract one another in millions of ways. Every query is reduced to a string of tokens and through massive computation and multiple feed back loops of say 50 or so layers, the next best token is produced stochastically from finding and converging to the parameters in memory that represent the best response to the query. The above is rough ? the curious should seek out the details. > >> I worked in the Stanford Knowledge Systems Lab at Stanford as well as for the DENDRAL project - AI in chemistry. My area of interest was using organic chemistry along with signal processing to analyze GCMS data with peak analysis gradients to detect inherited rare diseases in children from the fractions of these children?s blood. It worked and these were not heuristics. > > But were they systems that learned, and that detected features that they hadn't been told about in advance? I can?t answer the above for all expert systems. But that is not a requirement for AI. Also, LLMs do not KNOW anything. They build stochastic sentences from a fixed memory of parameters. They do NOT think. Rather they COMPUTE. There is NO ONE HOME. And they are explicitly trained to produce sentences that are pleasing to humans. One of the final phases of training is pumping through millions of queries and using RLHF - Reinforced Learning with Human Feedback to produce grammatically pretty and pleasing sentences. RLHF uses mathematical algorithms. I forget their names. Super cool ! This works for all languages, eg, Chinese tokens are characters - like ???????, ??etc. I?ll close with Yann LeCun?s description of LLMs and why he?s taking a divergent path to create models for AGI: LLMs regurgitate stochastic sentences ! Yann along with Geoffrey Hinton and Yoshua Benglo received the Turing prize for deep learning in 2018. > >> The spectra that were produced were give to several Phd?s in chemistry to identify the compounds, they were stored in a library for future recognition. Is it AI now? No. Was it then. Yes. What we did was state-of-the-art AI between 1970 and 1990 or so. >> Ed Feigenbaum who is a good friend received the Turing Prize along with Raj Redy for their work on expert systems in 1994. > > Yes, and I didn't mean to insult all that work (so I should have chosen my words more carefully). > >> I am currently writing a history of AI from 1955 to the present from the perspective that it was a Gestalt. Have looked into what was AI for each of the multiple versions over the years. I?ve been in computer science and math for 60 years, lived through these, epochs, and believe me there were battles noted in the press initiated by antagonists who were worried that expert systems were going to replace the experts. > > I never thought that, but it's already happening to a considerable extent with LLMs. > True - but better said - to help them work and do research more efficiently. This is true in math too - which I carefully follow. >> AI evolved over the years to where we are now and definitely NOT as a SILO. > > However, multi-layer perceptrons were basically ignored until a few years ago. (That includes me ignoring them when working on speech recognition in the late 1960s, because they were computationally infeasible.) > Exactly. We knew about neural nets but they were just toys back then. Everything was mega-x. Sorry for any of the typos above. Bye for now. When work is play for someone, all is well and this someone is a happy person (-: As my granddaughter always said, Papi (me) let?s play play play. Bill > Peace > Brian > >> Bill >> ?????????? >> ????? >> >> ???? ????????? >> ?????????? ????? ???? ????????? From b_a_denny at yahoo.com Tue Jun 23 09:52:24 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Tue, 23 Jun 2026 16:52:24 +0000 (UTC) Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: <580818707.1307788.1782233544689@mail.yahoo.com> As to books i would suggest reading the second edition of Richard Sutton's and Andrew Barto's book "Reinforcement Learning: An Introduction". They won the Turing award for the work last year.? I understand the book is very readable (no heavy math) but i haven't gotten around to reading my copy yet. By chance i also discovered a neuroscientist had done his thesis on this topic if you want to see what he did. "https://www.researchgate.net/publication/304048183_Human_Reinforcement_Learning_Insights_from_intracranial_recordings_and_stimulation"?https://mail.yahoo.com/n/search/keyword=martha&folders=2/messages/AOFV5jheakJdaAAUbgl14IuUswU?.src=ym&reason=myc#:~:text=https%3A//www.researchgate.net/publication/304048183_Human_Reinforcement_Learning_Insights_from_intracranial_recordings_and_stimulation barbara On Monday, June 22, 2026 at 07:54:28 PM PDT, Brian E Carpenter via Internet-history wrote: > Anybody else know more about the history of the use of AI *within* (not > just on top of) The Internet?? For example, when there are problems in > today's Internet, are AI techniques and tools used to diagnose and > repair them?? What's the History of such things? 1) For anyone interested in the history of AI, as well as how AI actually works, I recommend Anil Ananthaswamy's "Why Machines Learn" [1]. It's quite hard going if you want to follow all the maths, but it recapitualates the history of AI very well. 2) In answer to Jack's question, I think very little AI has been applied to networking in the past, but this is changing *very* fast. When we designed the GRASP protocol for autonomic networking, it was aimed at "intelligent" agents but we didn't have any - and that was only 5 years ago. Now the IETF is littered with drafts about intelligent agents. Big changes are coming. I have no doubt that 'expert systems' techniques have been applied, especially to network design, but expert systems were not real AI; they did incorporate heuristics, and anything using heuristics could be called AI at one stage. Not any more. A year ago I gave a talk about AI for network management [2] (aimed at final year students). I haven't checked recently, but I assume it's way out of date. [1] https://www.penguin.co.uk/books/446849/why-machines-learn-by-ananthaswamy-anil/9781802060867 [2] https://cs.auckland.ac.nz/~brian/CS389-BC-AI-netwk2025.pdf Regards/Ng? mihi ? ? Brian Carpenter On 23-Jun-26 07:50, Jack Haverty via Internet-history wrote: > With AI now in the news all the time, I've been wondering about the > history of AI in The Internet, i.e., how AI technology has been used > within the Internet over the years, rather than just a service which > people can access over The Internet.? ?Personally I only know of a few > such uses of AI in a communications role, quite a while ago. ?But maybe > others remember more...? > > In the mid-1970s, I was involved in an AI project at MIT in Licklider's > group, sponsored by ARPA.? ?It wasn't really part of the ARPANET but it > did relate to communications.? ?The project involved using AI techniques > of that era, then called "expert systems", to have computers decode > hand-sent Morse radio transmissions.? ?The Intelligence Community was > apparently quite interested in this problem at that time, and the > project ran for several years and was deemed a success. > > FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 for > info on that AI project.? Personally, I think of that 1990 report as an > enumeration of "ARPA's Greatest Hits".? ?There are six projects listed > from the Information Processing Techniques Office of ARPA (IPTO, where > all the networking and computer projects happened).? Another project of > those six IPTO successes is The ARPANET, described in Chapter 20.? ?The > Internet was probably still too immature to be included at the time . > > Similarly, while I was at BBN in the 1980s there were tools developed to > visualize activity in the ARPANET, and do tasks involved in network > design - figuring out where new lines were needed, reconfiguring the > topology of the ARPANET to address changes in traffic patterns, and > other such analyses.? I'm not sure those projects would be recognized as > "AI" today, but they were widely used to manage a variety of networks > such as the ARPANET and DDN. The idea was to use computers to augment > people skills, as it is in today's AI as well. > > Bob Kahn and I had a discussion at some point in late 1982 about AI and > networks in ARPA's world.? ?Bob was aware of the ARPA-sponsored project > done at MIT in the mid 1970s which used AI "expert system" techniques to > decode Morse code.? ?We mused that a similar expert system approach > could be used to manage networks, especially the emerging Internet which > was much more complex than the ARPANET and had few tools available for > operations and management. > > The concept was to use the plentiful supply of BBN engineers who had > been debugging Internet crises for several years as the "experts", > observing how they worked, and translating what they did into software > to do tasks "automatically".? ?That started the "Automated Network > Management" (ANM) project as a new research task at BBN.? ?I wrote the > proposal to ARPA and waited for the contract to be signed.? But when BBN > reorganized in July 1983 the ANM contract and I went separate ways so I > never got to work on ANM and I don't think there was enough detail in > the proposal to convey the concept onwards to the new project team. > > It strikes me that the real driver of AI today has evolved only partly > from advances in algorithms such as LLMs.? I think another major factor > has been the massive drop in computing costs over the last > half-century.? AI systems, such as that Morse Code project, were > affordable only by well-funded parts of government in the 1970s, > requiring millions of dollars/pounds/rubles/yen/lira to field a system > that provided minimal capability.? Today, the system that cost millions > in the 1970s now costs pennies by comparison. > > Anybody else know more about the history of the use of AI *within* (not > just on top of) The Internet?? ?For example, when there are problems in > today's Internet, are AI techniques and tools used to diagnose and > repair them?? ?What's the History of such things? > > /Jack Haverty > > -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From marc at petit-huguenin.org Tue Jun 23 10:12:31 2026 From: marc at petit-huguenin.org (Marc Petit-Huguenin) Date: Tue, 23 Jun 2026 10:12:31 -0700 Subject: [ih] History of AI and Internet In-Reply-To: <55D00418-C37F-4DCD-98C6-1B7870C41525@comcast.net> References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> <55D00418-C37F-4DCD-98C6-1B7870C41525@comcast.net> Message-ID: <36d043ed-4cf8-440a-b6a7-488c265a9b2c@petit-huguenin.org> I recently found in a DARPA solicitation a nice, concise, terminology for these two categories: Machine Learning (ML) for the first one, Automated Reasoning (AR) for the second. At this point in time, AI as a standalone initialism is a marketing term for the former, and used mostly to separate people from their money and their job. To go back to Internet history, I'd like to point out that not only AI/AR has a long history, but it is very much alive in the form of rules that can be applied to the semantic web in particular, and knowledge graphs in general (E.g., RuleML and ErgoAI). On 6/23/26 6:52 AM, John Day via Internet-history wrote: > I was told once that there are two forms of AI: > 1) Those that are trying to understand and then emulate how the brain works, and > 2) Those that are just trying to create something that acts like the brain. (I guess pass the Turing Test, but we are way beyond that.) > > Are we talking about the second case as AI? > > Is there any real progress on the first? > > Take care, > John > >> On Jun 23, 2026, at 02:10, Brian E Carpenter via Internet-history wrote: >> >> On 23-Jun-26 17:25, william yeager wrote: >>>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history > wrote: >>>> >>>> but expert systems were not real AI >>> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. >> >> Fair enough, but all the expert systems I saw were basically rules created by human experts. >> >>> I worked in the Stanford Knowledge Systems Lab at Stanford as well as for the DENDRAL project - AI in chemistry. My area of interest was using organic chemistry along with signal processing to analyze GCMS data with peak analysis gradients to detect inherited rare diseases in children from the fractions of these children?s blood. It worked and these were not heuristics. >> >> But were they systems that learned, and that detected features that they hadn't been told about in advance? >> >>> The spectra that were produced were give to several Phd?s in chemistry to identify the compounds, they were stored in a library for future recognition. Is it AI now? No. Was it then. Yes. What we did was state-of-the-art AI between 1970 and 1990 or so. >>> Ed Feigenbaum who is a good friend received the Turing Prize along with Raj Redy for their work on expert systems in 1994. >> >> Yes, and I didn't mean to insult all that work (so I should have chosen my words more carefully). >> >>> I am currently writing a history of AI from 1955 to the present from the perspective that it was a Gestalt. Have looked into what was AI for each of the multiple versions over the years. I?ve been in computer science and math for 60 years, lived through these, epochs, and believe me there were battles noted in the press initiated by antagonists who were worried that expert systems were going to replace the experts. >> >> I never thought that, but it's already happening to a considerable extent with LLMs. >> >>> AI evolved over the years to where we are now and definitely NOT as a SILO. >> >> However, multi-layer perceptrons were basically ignored until a few years ago. (That includes me ignoring them when working on speech recognition in the late 1960s, because they were computationally infeasible.) >> >> Peace >> Brian >> >>> Bill >>> ?????????? >>> ????? >>> >>> ???? ????????? >>> >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- Marc Petit-Huguenin Email: marc at petit-huguenin.org Blog: https://medium.com/@petithug Profile: https://www.linkedin.com/in/petithug -------------- next part -------------- A non-text attachment was scrubbed... Name: OpenPGP_signature.asc Type: application/pgp-signature Size: 840 bytes Desc: OpenPGP digital signature URL: From johnl at iecc.com Tue Jun 23 10:34:51 2026 From: johnl at iecc.com (John Levine) Date: 23 Jun 2026 13:34:51 -0400 Subject: [ih] History of AI and Internet In-Reply-To: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> Message-ID: <20260623173451.E18F411178751@ary.qy> It appears that Brian E Carpenter via Internet-history said: >On 23-Jun-26 17:25, william yeager wrote: >> >>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history > wrote: >>> >>> but expert systems were not real AI >> >> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. > >Fair enough, but all the expert systems I saw were basically rules created by human experts. Well, yeah, that was how people did AI when I was in school in the 1970s. Try to come up with more or less realistic models of the way people think. I think everyone was surprised when Google fed the parallel English and French texts of the Canadian Hansard into a machine learning system, and got a much better translator than any of the existing ones that were built with explicit language models. Now most AI is variations of dump in a lot of data and let the algorithms turn it into whatever model they get, but that's a huge change in mindset from AI 50 years ago. R's, John From b_a_denny at yahoo.com Tue Jun 23 12:52:09 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Tue, 23 Jun 2026 19:52:09 +0000 (UTC) Subject: [ih] History of AI and Internet In-Reply-To: <580818707.1307788.1782233544689@mail.yahoo.com> References: <580818707.1307788.1782233544689@mail.yahoo.com> Message-ID: <1942935820.1377156.1782244329302@mail.yahoo.com> Big oops , please ignore the second link to yahoo mail.? Somehow my? request to copy link from my old email? on the topic got created when i was trying just to retrieve the reference.? I didn't know yahoo? mail would even do something like that!!! barbara On Tuesday, June 23, 2026 at 09:52:24 AM PDT, Barbara Denny wrote: As to books i would suggest reading the second edition of Richard Sutton's and Andrew Barto's book "Reinforcement Learning: An Introduction". They won the Turing award for the work last year.? I understand the book is very readable (no heavy math) but i haven't gotten around to reading my copy yet. By chance i also discovered a neuroscientist had done his thesis on this topic if you want to see what he did. "https://www.researchgate.net/publication/304048183_Human_Reinforcement_Learning_Insights_from_intracranial_recordings_and_stimulation"?https://mail.yahoo.com/n/search/keyword=martha&folders=2/messages/AOFV5jheakJdaAAUbgl14IuUswU?.src=ym&reason=myc#:~:text=https%3A//www.researchgate.net/publication/304048183_Human_Reinforcement_Learning_Insights_from_intracranial_recordings_and_stimulation barbara On Monday, June 22, 2026 at 07:54:28 PM PDT, Brian E Carpenter via Internet-history wrote: > Anybody else know more about the history of the use of AI *within* (not > just on top of) The Internet?? For example, when there are problems in > today's Internet, are AI techniques and tools used to diagnose and > repair them?? What's the History of such things? 1) For anyone interested in the history of AI, as well as how AI actually works, I recommend Anil Ananthaswamy's "Why Machines Learn" [1]. It's quite hard going if you want to follow all the maths, but it recapitualates the history of AI very well. 2) In answer to Jack's question, I think very little AI has been applied to networking in the past, but this is changing *very* fast. When we designed the GRASP protocol for autonomic networking, it was aimed at "intelligent" agents but we didn't have any - and that was only 5 years ago. Now the IETF is littered with drafts about intelligent agents. Big changes are coming. I have no doubt that 'expert systems' techniques have been applied, especially to network design, but expert systems were not real AI; they did incorporate heuristics, and anything using heuristics could be called AI at one stage. Not any more. A year ago I gave a talk about AI for network management [2] (aimed at final year students). I haven't checked recently, but I assume it's way out of date. [1] https://www.penguin.co.uk/books/446849/why-machines-learn-by-ananthaswamy-anil/9781802060867 [2] https://cs.auckland.ac.nz/~brian/CS389-BC-AI-netwk2025.pdf Regards/Ng? mihi ? ? Brian Carpenter On 23-Jun-26 07:50, Jack Haverty via Internet-history wrote: > With AI now in the news all the time, I've been wondering about the > history of AI in The Internet, i.e., how AI technology has been used > within the Internet over the years, rather than just a service which > people can access over The Internet.? ?Personally I only know of a few > such uses of AI in a communications role, quite a while ago. ?But maybe > others remember more...? > > In the mid-1970s, I was involved in an AI project at MIT in Licklider's > group, sponsored by ARPA.? ?It wasn't really part of the ARPANET but it > did relate to communications.? ?The project involved using AI techniques > of that era, then called "expert systems", to have computers decode > hand-sent Morse radio transmissions.? ?The Intelligence Community was > apparently quite interested in this problem at that time, and the > project ran for several years and was deemed a success. > > FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 for > info on that AI project.? Personally, I think of that 1990 report as an > enumeration of "ARPA's Greatest Hits".? ?There are six projects listed > from the Information Processing Techniques Office of ARPA (IPTO, where > all the networking and computer projects happened).? Another project of > those six IPTO successes is The ARPANET, described in Chapter 20.? ?The > Internet was probably still too immature to be included at the time . > > Similarly, while I was at BBN in the 1980s there were tools developed to > visualize activity in the ARPANET, and do tasks involved in network > design - figuring out where new lines were needed, reconfiguring the > topology of the ARPANET to address changes in traffic patterns, and > other such analyses.? I'm not sure those projects would be recognized as > "AI" today, but they were widely used to manage a variety of networks > such as the ARPANET and DDN. The idea was to use computers to augment > people skills, as it is in today's AI as well. > > Bob Kahn and I had a discussion at some point in late 1982 about AI and > networks in ARPA's world.? ?Bob was aware of the ARPA-sponsored project > done at MIT in the mid 1970s which used AI "expert system" techniques to > decode Morse code.? ?We mused that a similar expert system approach > could be used to manage networks, especially the emerging Internet which > was much more complex than the ARPANET and had few tools available for > operations and management. > > The concept was to use the plentiful supply of BBN engineers who had > been debugging Internet crises for several years as the "experts", > observing how they worked, and translating what they did into software > to do tasks "automatically".? ?That started the "Automated Network > Management" (ANM) project as a new research task at BBN.? ?I wrote the > proposal to ARPA and waited for the contract to be signed.? But when BBN > reorganized in July 1983 the ANM contract and I went separate ways so I > never got to work on ANM and I don't think there was enough detail in > the proposal to convey the concept onwards to the new project team. > > It strikes me that the real driver of AI today has evolved only partly > from advances in algorithms such as LLMs.? I think another major factor > has been the massive drop in computing costs over the last > half-century.? AI systems, such as that Morse Code project, were > affordable only by well-funded parts of government in the 1970s, > requiring millions of dollars/pounds/rubles/yen/lira to field a system > that provided minimal capability.? Today, the system that cost millions > in the 1970s now costs pennies by comparison. > > Anybody else know more about the history of the use of AI *within* (not > just on top of) The Internet?? ?For example, when there are problems in > today's Internet, are AI techniques and tools used to diagnose and > repair them?? ?What's the History of such things? > > /Jack Haverty > > -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From jeanjour at comcast.net Tue Jun 23 13:09:00 2026 From: jeanjour at comcast.net (John Day) Date: Tue, 23 Jun 2026 16:09:00 -0400 Subject: [ih] History of AI and Internet In-Reply-To: <20260623173451.E18F411178751@ary.qy> References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> <20260623173451.E18F411178751@ary.qy> Message-ID: I know this is naive and idealistic, but what do we learn about models of languages from such exercises? Is this just engineering? Or do we learn something about how the brain works? It could be. John > On Jun 23, 2026, at 13:34, John Levine via Internet-history wrote: > > It appears that Brian E Carpenter via Internet-history said: >> On 23-Jun-26 17:25, william yeager wrote: >>> >>>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history > > wrote: >>>> >>>> but expert systems were not real AI >>> >>> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. >> >> Fair enough, but all the expert systems I saw were basically rules created by human experts. > > Well, yeah, that was how people did AI when I was in school in the 1970s. Try to > come up with more or less realistic models of the way people think. > > I think everyone was surprised when Google fed the parallel English and French > texts of the Canadian Hansard into a machine learning system, and got a much > better translator than any of the existing ones that were built with explicit > language models. Now most AI is variations of dump in a lot of data and let the > algorithms turn it into whatever model they get, but that's a huge change in > mindset from AI 50 years ago. > > R's, > John > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From dwaitzman at gmail.com Tue Jun 23 13:40:01 2026 From: dwaitzman at gmail.com (D Waitzman) Date: Tue, 23 Jun 2026 16:40:01 -0400 Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: <61BBDD20-1380-4173-94B5-34D07764D204@gmail.com> Jack Haverty > wrote: > That started the "Automated Network Management" (ANM) project as a new research task at BBN I worked on ANM from 1989 to 1997, though it had been renamed Advanced Network Management (still ANM) before I started. I was told that the team realized that it was not viable to be 'Automated'. The only remnant of AI was that the UI was written in Lisp (Common Lisp, with Flavors IIRC). I took Professor Marvin Minsky's Society of Minds class (6.868) at MIT in 1992. Early this year, in one of my first attempts at using Gemini "deep research", I asked it what Professor Marvin Minsky would have thought about today's AI systems. Its answer was the typical on-the-surface well-written piece you'd expect but I did not do an in-depth analysis. I suspect that Minsky would be strongly on the side of "today's AI is not real AGI". And regarding rule-based systems and the Internet -- how many of you worked on DEC VAXen whose final configuration was done by DEC's R-1 expert system? (See https://cdn.aaai.org/AAAI/1980/AAAI80-076.pdf) -- David Waitzman From byeager at fastmail.fm Tue Jun 23 14:04:12 2026 From: byeager at fastmail.fm (william yeager) Date: Tue, 23 Jun 2026 14:04:12 -0700 Subject: [ih] History of AI and Internet In-Reply-To: <61BBDD20-1380-4173-94B5-34D07764D204@gmail.com> References: <61BBDD20-1380-4173-94B5-34D07764D204@gmail.com> Message-ID: Hi? > On Jun 23, 2026, at 1:40 PM, D Waitzman via Internet-history wrote: > > > Jack Haverty > wrote: >> That started the "Automated Network Management" (ANM) project as a new research task at BBN > > I worked on ANM from 1989 to 1997, though it had been renamed Advanced Network Management (still ANM) before I started. I was told that the team realized that it was not viable to be 'Automated'. The only remnant of AI was that the UI was written in Lisp (Common Lisp, with Flavors IIRC). > > > > I took Professor Marvin Minsky's Society of Minds class (6.868) at MIT in 1992. Early this year, in one of my first attempts at using Gemini "deep research", I asked it what Professor Marvin Minsky would have thought about today's AI systems. Its answer was the typical on-the-surface well-written piece you'd expect but I did not do an in-depth analysis. I suspect that Minsky would be strongly on the side of "today's AI is not real AGI?. > Dr. Marvin Minsky was one of the co-signers of John McCarthy?s 1955 document coining the name AI and defining what he thought it should be. This is before he became a co-founder of the MIT AI-lab. Super wonderful, brilliant fellow. I agree with you. By the way, while at Stanford, my research to use AI techniques to monitor and manage the Stanford local area ethernet using the SNMP data of hosts and routers, Bayesian Statistics, straight forward statistics, the laboratory's calendar of events, build and maintain a network topology internal map looking for changes that would trigger an investigation, keep track of my labs network traffic 24x7 looking for out-of-band abnormalities that would also trigger an alert , etc., and was making decent progress when DEC was having financial problems 1990 and cut my financial support. Nice learning experience. > > And regarding rule-based systems and the Internet -- how many of you worked on DEC VAXen whose final configuration was done by DEC's R-1 expert system? (See https://cdn.aaai.org/AAAI/1980/AAAI80-076.pdf) > > > -- > David Waitzman > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history ?????????? ????? ???? ????????? From b_a_denny at yahoo.com Tue Jun 23 14:08:35 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Tue, 23 Jun 2026 21:08:35 +0000 (UTC) Subject: [ih] History of AI and Internet In-Reply-To: <61BBDD20-1380-4173-94B5-34D07764D204@gmail.com> References: <61BBDD20-1380-4173-94B5-34D07764D204@gmail.com> Message-ID: <610754579.1424702.1782248915235@mail.yahoo.com> Have you heard of bothunter? I tried it out in the lab at Boeing (2010).? I am not sure whether people will call it AI but i thought it was interesting for its time. https://www.csl.sri.com/~vinod/papers/bothunter.pdf barbara On Tuesday, June 23, 2026 at 01:40:22 PM PDT, D Waitzman via Internet-history wrote: Jack Haverty > wrote: > That started the "Automated Network Management" (ANM) project as a new research task at BBN I worked on ANM from 1989 to 1997, though it had been renamed Advanced Network Management (still ANM) before I started.? I was told that the team realized that it was not viable to be 'Automated'.? The only remnant of AI was that the UI was written in Lisp (Common Lisp, with Flavors IIRC). I took Professor Marvin Minsky's Society of Minds class (6.868) at MIT in 1992. Early this year, in one of my first attempts at using Gemini "deep research", I asked it what Professor Marvin Minsky would have thought about today's AI systems.? Its answer was the typical on-the-surface well-written piece you'd expect but I did not do an in-depth analysis.? I suspect that Minsky would be strongly on the side of "today's AI is not real AGI". And regarding rule-based systems and the Internet -- how many of you worked on DEC VAXen whose final configuration was done by DEC's R-1 expert system?? (See https://cdn.aaai.org/AAAI/1980/AAAI80-076.pdf) -- David Waitzman -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From brian.e.carpenter at gmail.com Tue Jun 23 14:14:50 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Wed, 24 Jun 2026 09:14:50 +1200 Subject: [ih] History of AI and Internet In-Reply-To: <93F30632-2721-44B5-852D-8DA3369212BF@fastmail.fm> References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> <93F30632-2721-44B5-852D-8DA3369212BF@fastmail.fm> Message-ID: <37c7216d-5b10-43d1-b11b-211de151ffb8@gmail.com> On 24-Jun-26 03:10, william yeager wrote: > Don?t want to spend too much time on this but apparently you?ve never read John McCarthy?s proposal for naming AI and what AI should do. Here it is: > > We propose that a 2 month, 10 manstudy of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire > > The study is to proceed on the basis ofthe conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulateit. Oh, I am quite familiar with that, and coincidentally I am just reading Matthew Cobb's new biography of Francis Crick, which brings out very clearly that Crick's work on brain function was very clearly contrasted with AI simulation. I suspect that Crick would have violently objected to McCarthy?s conjecture. (Crick wasn't interested in artificial neural networks, because real neurons don't have a backward propagation mechanism.) Regards/Ng? mihi Brian Carpenter >> On Jun 22, 2026, at 11:10 PM, Brian E Carpenter wrote: >> >> On 23-Jun-26 17:25, william yeager wrote: >>>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history >> wrote: >>>> >>>> but expert systems were not real AI >>> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. >> >> Fair enough, but all the expert systems I saw were basically rules created by human experts. > > This is true and the Expert Systems did simulate human thinking. These were not simple rules. Ted Shortliffe?s MYCIN and EMYCIN incorporated rules for diagnosing and prescribing antibiotics. These systems did as well as 90% of physicians. > > I personally worked on T-Helper that was used to remotely diagnose and treat AIDs patients and could be used by rural communities that lacked the expertise. The ideas weren?t mine. I helped a Phd student in medical informatics - ie - the student was an MD and wanted to add a Phd in the computational aspect of medicine - the underlying code misused TCP/IP - the latter was buried in the bowels of the large system and as I dug into it I also began to underside the medical side. Amazing software. > > The work in Expert Systems covered so many bases and Stanford the a national resource for the work in experimentation in AI and Medicine in the 1970s-90s. > > Yes, Expert Systems did simulate human thinking. This was not about generating stochastic sentences like LLMs do. I really understand the internals of these systems. For me the math was easy since I did my Phd studies in math at U of Washington. I am amazed at the beauty in LLMs and with the arrival in 2017 of the Foundation that enabled deep learning to scale and compute on 20 thousand NVIDIA GPUs in parallel; multiplying token based, weighted matrices on the order of 4096x4096 or larger along with smaller queries, keys and values matrices all created during training, the the embedded matrix 80000x4096 tokens + a few more ? the details are a thing of pure mathematical wonder to me. Do note that if one understands ?and not just knows advanced multivariate calculus and linear algebra along with standard error calculations and a few other details, they ought to be able to figure out how LLMs work. WHY is more complicated and worth another long discussion. > > The basic thing to understand about LLMs is that they are 90% training and 10% execution. They have a fixed memory of about 30 billion parameters and are becoming larger. The do not keep the vast amount of text that is scrapped off of the Internet; They have a fixed vocabulary of about 60 thousand words; and about 80 thousand tokens - the embedded array - on which all computation is based. Tokens are like syllables - pieces of words - have binary values that map to the text. So Bill might be ?bi? and ?ll?. One knows a token by the friends it keeps. They attract one another in millions of ways. Every query is reduced to a string of tokens and through massive computation and multiple feed back loops of say 50 or so layers, the next best token is produced stochastically from finding and converging to the parameters in memory that represent the best response to the query. > > The above is rough ? the curious should seek out the details. >> >>> I worked in the Stanford Knowledge Systems Lab at Stanford as well as for the DENDRAL project - AI in chemistry. My area of interest was using organic chemistry along with signal processing to analyze GCMS data with peak analysis gradients to detect inherited rare diseases in children from the fractions of these children?s blood. It worked and these were not heuristics. >> >> But were they systems that learned, and that detected features that they hadn't been told about in advance? > > I can?t answer the above for all expert systems. But that is not a requirement for AI. Also, LLMs do not KNOW anything. They build stochastic sentences from a fixed memory of parameters. They do NOT think. Rather they COMPUTE. There is NO ONE HOME. And they are explicitly trained to produce sentences that are pleasing to humans. One of the final phases of training is pumping through millions of queries and using RLHF - Reinforced Learning with Human Feedback to produce grammatically pretty and pleasing sentences. RLHF uses mathematical algorithms. I forget their names. Super cool ! This works for all languages, eg, Chinese tokens are characters - like ???????, ??etc. > > I?ll close with Yann LeCun?s description of LLMs and why he?s taking a divergent path to create models for AGI: LLMs regurgitate stochastic sentences ! Yann along with Geoffrey Hinton and Yoshua Benglo received the Turing prize for deep learning in 2018. > > >> >>> The spectra that were produced were give to several Phd?s in chemistry to identify the compounds, they were stored in a library for future recognition. Is it AI now? No. Was it then. Yes. What we did was state-of-the-art AI between 1970 and 1990 or so. >>> Ed Feigenbaum who is a good friend received the Turing Prize along with Raj Redy for their work on expert systems in 1994. >> >> Yes, and I didn't mean to insult all that work (so I should have chosen my words more carefully). >> >>> I am currently writing a history of AI from 1955 to the present from the perspective that it was a Gestalt. Have looked into what was AI for each of the multiple versions over the years. I?ve been in computer science and math for 60 years, lived through these, epochs, and believe me there were battles noted in the press initiated by antagonists who were worried that expert systems were going to replace the experts. >> >> I never thought that, but it's already happening to a considerable extent with LLMs. >> > True - but better said - to help them work and do research more efficiently. > This is true in math too - which I carefully follow. > >>> AI evolved over the years to where we are now and definitely NOT as a SILO. >> >> However, multi-layer perceptrons were basically ignored until a few years ago. (That includes me ignoring them when working on speech recognition in the late 1960s, because they were computationally infeasible.) >> > > Exactly. We knew about neural nets but they were just toys back then. Everything was mega-x. > > Sorry for any of the typos above. > > Bye for now. When work is play for someone, all is well and this someone is a happy person (-: As my granddaughter always said, Papi (me) let?s play play play. > > Bill >> Peace >> ??Brian >> >>> Bill >>> ?????????? >>> ????? >>> >>> ???? ????????? >>> > > ?????????? > ????? > > ???? ????????? > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > From byeager at fastmail.fm Tue Jun 23 14:31:35 2026 From: byeager at fastmail.fm (william yeager) Date: Tue, 23 Jun 2026 14:31:35 -0700 Subject: [ih] History of AI and Internet In-Reply-To: <37c7216d-5b10-43d1-b11b-211de151ffb8@gmail.com> References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> <93F30632-2721-44B5-852D-8DA3369212BF@fastmail.fm> <37c7216d-5b10-43d1-b11b-211de151ffb8@gmail.com> Message-ID: <8FF0077F-7B16-4F37-8471-C641836A3CCB@fastmail.fm> LOL ! Back propogation is extremely important to assure that as computation takes proceeds on successive layers, that the current layers predecessors are informed and correct what is learned about the current token in R^n at the current layer. N is called the dimension of the LLM, say, 4096. Extremely important. I personally think the neuron is a misnomer. It is a mathematically defined relationship between active tokens and the memory. But, an LLM does definitely use a training created, digital version of the human brain. I?d forget the differences between real neurons and an LLM?s. Just note that the human brain has an associative memory and a LLM does not. The later also operates on an average of 20 watts ! Now that is astounding. I think we also use tokens in the sense that if I want to remember a forgotten name, I just go a, b, c, ?, z over and over and various names arrive, and with luck the desired one. I?m sure most of us have done this. > On Jun 23, 2026, at 2:14 PM, Brian E Carpenter wrote: > > On 24-Jun-26 03:10, william yeager wrote: >> Don?t want to spend too much time on this but apparently you?ve never read John McCarthy?s proposal for naming AI and what AI should do. Here it is: >> >> We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire >> >> The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. > Oh, I am quite familiar with that, and coincidentally I am just reading Matthew Cobb's new biography of Francis Crick, which brings out very clearly that Crick's work on brain function was very clearly contrasted with AI simulation. I suspect that Crick would have violently objected to McCarthy?s conjecture. (Crick wasn't interested in artificial neural networks, because real neurons don't have a backward propagation mechanism.) > > Regards/Ng? mihi > Brian Carpenter >>> On Jun 22, 2026, at 11:10 PM, Brian E Carpenter > wrote: >>> >>> On 23-Jun-26 17:25, william yeager wrote: >>>>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history >> wrote: >>>>> >>>>> but expert systems were not real AI >>>> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. >>> >>> Fair enough, but all the expert systems I saw were basically rules created by human experts. >> >> This is true and the Expert Systems did simulate human thinking. These were not simple rules. Ted Shortliffe?s MYCIN and EMYCIN incorporated rules for diagnosing and prescribing antibiotics. These systems did as well as 90% of physicians. >> >> I personally worked on T-Helper that was used to remotely diagnose and treat AIDs patients and could be used by rural communities that lacked the expertise. The ideas weren?t mine. I helped a Phd student in medical informatics - ie - the student was an MD and wanted to add a Phd in the computational aspect of medicine - the underlying code misused TCP/IP - the latter was buried in the bowels of the large system and as I dug into it I also began to underside the medical side. Amazing software. >> >> The work in Expert Systems covered so many bases and Stanford the a national resource for the work in experimentation in AI and Medicine in the 1970s-90s. >> >> Yes, Expert Systems did simulate human thinking. This was not about generating stochastic sentences like LLMs do. I really understand the internals of these systems. For me the math was easy since I did my Phd studies in math at U of Washington. I am amazed at the beauty in LLMs and with the arrival in 2017 of the Foundation that enabled deep learning to scale and compute on 20 thousand NVIDIA GPUs in parallel; multiplying token based, weighted matrices on the order of 4096x4096 or larger along with smaller queries, keys and values matrices all created during training, the the embedded matrix 80000x4096 tokens + a few more ? the details are a thing of pure mathematical wonder to me. Do note that if one understands and not just knows advanced multivariate calculus and linear algebra along with standard error calculations and a few other details, they ought to be able to figure out how LLMs work. WHY is more complicated and worth another long discussion. >> >> The basic thing to understand about LLMs is that they are 90% training and 10% execution. They have a fixed memory of about 30 billion parameters and are becoming larger. The do not keep the vast amount of text that is scrapped off of the Internet; They have a fixed vocabulary of about 60 thousand words; and about 80 thousand tokens - the embedded array - on which all computation is based. Tokens are like syllables - pieces of words - have binary values that map to the text. So Bill might be ?bi? and ?ll?. One knows a token by the friends it keeps. They attract one another in millions of ways. Every query is reduced to a string of tokens and through massive computation and multiple feed back loops of say 50 or so layers, the next best token is produced stochastically from finding and converging to the parameters in memory that represent the best response to the query. >> >> The above is rough ? the curious should seek out the details. >>> >>>> I worked in the Stanford Knowledge Systems Lab at Stanford as well as for the DENDRAL project - AI in chemistry. My area of interest was using organic chemistry along with signal processing to analyze GCMS data with peak analysis gradients to detect inherited rare diseases in children from the fractions of these children?s blood. It worked and these were not heuristics. >>> >>> But were they systems that learned, and that detected features that they hadn't been told about in advance? >> >> I can?t answer the above for all expert systems. But that is not a requirement for AI. Also, LLMs do not KNOW anything. They build stochastic sentences from a fixed memory of parameters. They do NOT think. Rather they COMPUTE. There is NO ONE HOME. And they are explicitly trained to produce sentences that are pleasing to humans. One of the final phases of training is pumping through millions of queries and using RLHF - Reinforced Learning with Human Feedback to produce grammatically pretty and pleasing sentences. RLHF uses mathematical algorithms. I forget their names. Super cool ! This works for all languages, eg, Chinese tokens are characters - like ???????, ??etc. >> >> I?ll close with Yann LeCun?s description of LLMs and why he?s taking a divergent path to create models for AGI: LLMs regurgitate stochastic sentences ! Yann along with Geoffrey Hinton and Yoshua Benglo received the Turing prize for deep learning in 2018. >> >> >>> >>>> The spectra that were produced were give to several Phd?s in chemistry to identify the compounds, they were stored in a library for future recognition. Is it AI now? No. Was it then. Yes. What we did was state-of-the-art AI between 1970 and 1990 or so. >>>> Ed Feigenbaum who is a good friend received the Turing Prize along with Raj Redy for their work on expert systems in 1994. >>> >>> Yes, and I didn't mean to insult all that work (so I should have chosen my words more carefully). >>> >>>> I am currently writing a history of AI from 1955 to the present from the perspective that it was a Gestalt. Have looked into what was AI for each of the multiple versions over the years. I?ve been in computer science and math for 60 years, lived through these, epochs, and believe me there were battles noted in the press initiated by antagonists who were worried that expert systems were going to replace the experts. >>> >>> I never thought that, but it's already happening to a considerable extent with LLMs. >>> >> True - but better said - to help them work and do research more efficiently. >> >> This is true in math too - which I carefully follow. >> >>>> AI evolved over the years to where we are now and definitely NOT as a SILO. >>> >>> However, multi-layer perceptrons were basically ignored until a few years ago. (That includes me ignoring them when working on speech recognition in the late 1960s, because they were computationally infeasible.) >>> >> >> Exactly. We knew about neural nets but they were just toys back then. Everything was mega-x. >> >> Sorry for any of the typos above. >> >> Bye for now. When work is play for someone, all is well and this someone is a happy person (-: As my granddaughter always said, Papi (me) let?s play play play. >> >> Bill >>> Peace >>> Brian >>> >>>> Bill >>>> ?????????? >>>> ????? >>>> >>>> ???? ????????? >>>> >> >> ?????????? >> ????? >> >> ???? ????????? >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> ?????????? ????? ???? ????????? From johnl at iecc.com Tue Jun 23 17:03:20 2026 From: johnl at iecc.com (John R. Levine) Date: 23 Jun 2026 20:03:20 -0400 Subject: [ih] History of AI and Internet In-Reply-To: References: <5c854014-c372-4d60-8998-d18a4b334e34@gmail.com> <20260623173451.E18F411178751@ary.qy> Message-ID: <6b267945-fcbd-3ecc-aec0-f417f7420ae3@iecc.com> > I know this is naive and idealistic, but what do we learn about models of languages from such exercises? > Is this just engineering? Or do we learn something about how the brain works? Real answer: nobody knows. Slightly longer answer: LLMs are a weighted function of billions, maybe trillions of weights. It's so many that as far as I know there is no way to tell what the model really is other than poking at it from the outside. People have built neural networks with threshold gates that operate sort of like biological neurons, but I don't see any reason to believe that the way they interconnect is like the way actual neurons do. R's, John >> On Jun 23, 2026, at 13:34, John Levine via Internet-history wrote: >> >> It appears that Brian E Carpenter via Internet-history said: >>> On 23-Jun-26 17:25, william yeager wrote: >>>> >>>>> On Jun 22, 2026, at 7:54 PM, Brian E Carpenter via Internet-history >> > wrote: >>>>> >>>>> but expert systems were not real AI >>>> >>>> Rubbish ! (Steven Hawking?s favorite denial). They WERE real AI at the time they existed. >>> >>> Fair enough, but all the expert systems I saw were basically rules created by human experts. >> >> Well, yeah, that was how people did AI when I was in school in the 1970s. Try to >> come up with more or less realistic models of the way people think. >> >> I think everyone was surprised when Google fed the parallel English and French >> texts of the Canadian Hansard into a machine learning system, and got a much >> better translator than any of the existing ones that were built with explicit >> language models. Now most AI is variations of dump in a lot of data and let the >> algorithms turn it into whatever model they get, but that's a huge change in >> mindset from AI 50 years ago. From mfidelman at protocoltechnologiesgroup.com Tue Jun 23 17:11:00 2026 From: mfidelman at protocoltechnologiesgroup.com (Miles Fidelman) Date: Wed, 24 Jun 2026 00:11:00 +0000 Subject: [ih] History of AI and Internet In-Reply-To: <610754579.1424702.1782248915235@mail.yahoo.com> References: <61BBDD20-1380-4173-94B5-34D07764D204@gmail.com> <610754579.1424702.1782248915235@mail.yahoo.com> Message-ID: Now, more than ever, we need anti-bot search & destroy technology. Sent from Samsung Galaxy smartphone. Get Outlook for Android ________________________________ From: Internet-history on behalf of Barbara Denny via Internet-history Sent: Tuesday, June 23, 2026 5:08:53 PM To: Internet-history Subject: Re: [ih] History of AI and Internet Have you heard of bothunter? I tried it out in the lab at Boeing (2010). I am not sure whether people will call it AI but i thought it was interesting for its time. https://www.csl.sri.com/~vinod/papers/bothunter.pdf barbara On Tuesday, June 23, 2026 at 01:40:22 PM PDT, D Waitzman via Internet-history wrote: Jack Haverty > wrote: > That started the "Automated Network Management" (ANM) project as a new research task at BBN I worked on ANM from 1989 to 1997, though it had been renamed Advanced Network Management (still ANM) before I started. I was told that the team realized that it was not viable to be 'Automated'. The only remnant of AI was that the UI was written in Lisp (Common Lisp, with Flavors IIRC). I took Professor Marvin Minsky's Society of Minds class (6.868) at MIT in 1992. Early this year, in one of my first attempts at using Gemini "deep research", I asked it what Professor Marvin Minsky would have thought about today's AI systems. Its answer was the typical on-the-surface well-written piece you'd expect but I did not do an in-depth analysis. I suspect that Minsky would be strongly on the side of "today's AI is not real AGI". And regarding rule-based systems and the Internet -- how many of you worked on DEC VAXen whose final configuration was done by DEC's R-1 expert system? (See https://cdn.aaai.org/AAAI/1980/AAAI80-076.pdf) -- David Waitzman -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From gregskinner0 at icloud.com Wed Jun 24 00:29:09 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Wed, 24 Jun 2026 00:29:09 -0700 Subject: [ih] History of AI and Internet In-Reply-To: <580818707.1307788.1782233544689@mail.yahoo.com> References: <580818707.1307788.1782233544689@mail.yahoo.com> Message-ID: On Jun 23, 2026, at 9:52?AM, Barbara Denny via Internet-history wrote: > > As to books i would suggest reading the second edition of Richard Sutton's and Andrew Barto's book "Reinforcement Learning: An Introduction". They won the Turing award for the work last year. I understand the book is very readable (no heavy math) but i haven't gotten around to reading my copy yet. > By chance i also discovered a neuroscientist had done his thesis on this topic if you want to see what he did. > "https://www.researchgate.net/publication/304048183_Human_Reinforcement_Learning_Insights_from_intracranial_recordings_and_stimulation" > > barbara > I found a copy of Sutton and Barto?s text online. [1] (It is noted as a work in progress.) Just skimming through the book, it draws quite a bit on dynamic programming (DP). That might be challenging for those who don?t use DP regularly or haven?t yet learned it. I also noticed that Martha Steenstrup reviewed the text and provided feedback to the authors. For those not familiar with her, she made key contributions to policy-based routing and mobile ad-hoc networking. --gregbo [1] https://web.stanford.edu/class/psych209/Readings/SuttonBartoIPRLBook2ndEd.pdf From jjd at jjd.com Wed Jun 24 08:14:53 2026 From: jjd at jjd.com (James J Dempsey) Date: Wed, 24 Jun 2026 11:14:53 -0400 Subject: [ih] History of AI and Internet Message-ID: <1397644.1782314093@serenity.jjd.com> Jack Haverty wrote: > Similarly, while I was at BBN in the 1980s there were tools developed to > visualize activity in the ARPANET, and do tasks involved in network > design - figuring out where new lines were needed, reconfiguring the > topology of the ARPANET to address changes in traffic patterns, and > other such analyses.? I'm not sure those projects would be recognized as > "AI" today, but they were widely used to manage a variety of networks > such as the ARPANET and DDN. Hi Jack. I think you are talking about "Designer" or what was commercialized as "DESIGNet", an expert system built in Lisp than ran on Symbolics machines. As you say, we used it for ARPANET, MILNET, other government networks as well as many commercial X.25 networks that we built using the same ARPANET technology. There is a pretty good description of DESIGNet in this article from Data Communications: https://columbia.edu/~agb6/papers/DesigNet.pdf A funny DESIGNet AI-related story. We had a new VP who was being given a demo of DESIGNet by Jeff Mayersohn, who was head of the Network Analysis group at BBN. As it was promoted as an AI system, this VP, not understanding rule-based expert systems, was skeptical and wanted to what part of it was "AI". He kept asking during the demo "Where's the AI?" One of the tools in DESIGnet was to check to ensure whether the proposed topology was 2-connected -- i.e. could two links in the network go down and the network would remain connected, not bifurcated. Jeff was demonstrating this tool and the system printed out a message saying something like "Yes, your design is 2-connected. Strangely enough, it is also 3-connected." The VP asked "How did it know to say 'strangely enough'?" And without pausing, Jeff said "That's AI!" --Jim Dempsey-- From johngrahamalvord at gmail.com Wed Jun 24 10:26:19 2026 From: johngrahamalvord at gmail.com (John Alvord) Date: Wed, 24 Jun 2026 10:26:19 -0700 Subject: [ih] History of AI and Internet In-Reply-To: References: <61BBDD20-1380-4173-94B5-34D07764D204@gmail.com> <610754579.1424702.1782248915235@mail.yahoo.com> Message-ID: On Tue, Jun 23, 2026 at 17:11 Miles Fidelman via Internet-history < internet-history at elists.isoc.org> wrote: > Now, more than ever, we need anti-bot search & destroy technology. > > Sent from Samsung Galaxy smartphone. > Get Outlook for Android > > ________________________________ > I From: Internet-history on > behalf of Barbara Denny via Internet-history < > internet-history at elists.isoc.org> > Sent: Tuesday, June 23, 2026 5:08:53 PM > To: Internet-history > Subject: Re: [ih] History of AI and Internet > > Have you heard of bothunter? I tried it out in the lab at Boeing (2010). > I am not sure whether people will call it AI but i thought it was > interesting for its time. > https://www.csl.sri.com/~vinod/papers/bothunter.pdf > barbara > On Tuesday, June 23, 2026 at 01:40:22 PM PDT, D Waitzman via > Internet-history wrote: > > > Jack Haverty > wrote: > > That started the "Automated Network Management" (ANM) project as a new > research task at BBN > > I worked on ANM from 1989 to 1997, though it had been renamed Advanced > Network Management (still ANM) before I started. I was told that the team > realized that it was not viable to be 'Automated'. The only remnant of AI > was that the UI was written in Lisp (Common Lisp, with Flavors IIRC). > > > > I took Professor Marvin Minsky's Society of Minds class (6.868) at MIT in > 1992. Early this year, in one of my first attempts at using Gemini "deep > research", I asked it what Professor Marvin Minsky would have thought about > today's AI systems. Its answer was the typical on-the-surface well-written > piece you'd expect but I did not do an in-depth analysis. I suspect that > Minsky would be strongly on the side of "today's AI is not real AGI". > > > And regarding rule-based systems and the Internet -- how many of you > worked on DEC VAXen whose final configuration was done by DEC's R-1 expert > system? (See https://cdn.aaai.org/AAAI/1980/AAAI80-076.pdf) > > > -- > David Waitzman > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From b_a_denny at yahoo.com Wed Jun 24 11:45:21 2026 From: b_a_denny at yahoo.com (Barbara Denny) Date: Wed, 24 Jun 2026 18:45:21 +0000 (UTC) Subject: [ih] History of AI and Internet In-Reply-To: References: <580818707.1307788.1782233544689@mail.yahoo.com> Message-ID: <1677377271.1720268.1782326721278@mail.yahoo.com> Just to be clear, the second edition was published (2018).? I don't know about the draft you read online.? I got my copy on amazon. Too bad all the tech bookstores i knew about in the Bay Area closed along time? ago. barbara On Wednesday, June 24, 2026 at 12:29:23 AM PDT, Greg Skinner wrote: On Jun 23, 2026, at 9:52?AM, Barbara Denny via Internet-history wrote: As to books i would suggest reading the second edition of Richard Sutton's and Andrew Barto's book "Reinforcement Learning: An Introduction". They won the Turing award for the work last year.? I understand the book is very readable (no heavy math) but i haven't gotten around to reading my copy yet. By chance i also discovered a neuroscientist had done his thesis on this topic if you want to see what he did. "https://www.researchgate.net/publication/304048183_Human_Reinforcement_Learning_Insights_from_intracranial_recordings_and_stimulation"? barbara I found a copy of Sutton and Barto?s text online. [1] (It is noted as a work in progress.) Just skimming through the book, it draws quite a bit on dynamic programming (DP). ?That might be challenging for those who don?t use DP regularly or haven?t yet learned it. I also noticed that Martha Steenstrup reviewed the text and provided feedback to the authors. ?For those not familiar with her, she made key contributions to policy-based routing and mobile ad-hoc networking. --gregbo [1]?https://web.stanford.edu/class/psych209/Readings/SuttonBartoIPRLBook2ndEd.pdf From 6.internet at gmail.com Wed Jun 24 15:22:19 2026 From: 6.internet at gmail.com (Sivan) Date: Thu, 25 Jun 2026 03:52:19 +0530 Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: Dear Jack Haverty, Your question "For example, when there are problems in today's Internet, are AI techniques and tools used to diagnose and repair them? What's the History of such things?" is immensely interesting. Concerns about A.I. momentarily set aside, are there initiatives underway to positively use A.I. tools to "diagnose and repair" problems in the Internet? For example, using A.I. to scan for malware, bots, phishing and other forms of technical and non-technical Abuse? Or using A.I. to scan and detect barriers to network protocols such as vpns? Or even using A.I. to scan and detect non-human content and other forms of Abuse? Sivasubramanian Muthusamy sender On Tue, 23 Jun, 2026, 01:20 Jack Haverty via Internet-history, < internet-history at elists.isoc.org> wrote: > With AI now in the news all the time, I've been wondering about the > history of AI in The Internet, i.e., how AI technology has been used > within the Internet over the years, rather than just a service which > people can access over The Internet. Personally I only know of a few > such uses of AI in a communications role, quite a while ago. But maybe > others remember more...? > > In the mid-1970s, I was involved in an AI project at MIT in Licklider's > group, sponsored by ARPA. It wasn't really part of the ARPANET but it > did relate to communications. The project involved using AI techniques > of that era, then called "expert systems", to have computers decode > hand-sent Morse radio transmissions. The Intelligence Community was > apparently quite interested in this problem at that time, and the > project ran for several years and was deemed a success. > > FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 for > info on that AI project. Personally, I think of that 1990 report as an > enumeration of "ARPA's Greatest Hits". There are six projects listed > from the Information Processing Techniques Office of ARPA (IPTO, where > all the networking and computer projects happened). Another project of > those six IPTO successes is The ARPANET, described in Chapter 20. The > Internet was probably still too immature to be included at the time . > > Similarly, while I was at BBN in the 1980s there were tools developed to > visualize activity in the ARPANET, and do tasks involved in network > design - figuring out where new lines were needed, reconfiguring the > topology of the ARPANET to address changes in traffic patterns, and > other such analyses. I'm not sure those projects would be recognized as > "AI" today, but they were widely used to manage a variety of networks > such as the ARPANET and DDN. The idea was to use computers to augment > people skills, as it is in today's AI as well. > > Bob Kahn and I had a discussion at some point in late 1982 about AI and > networks in ARPA's world. Bob was aware of the ARPA-sponsored project > done at MIT in the mid 1970s which used AI "expert system" techniques to > decode Morse code. We mused that a similar expert system approach > could be used to manage networks, especially the emerging Internet which > was much more complex than the ARPANET and had few tools available for > operations and management. > > The concept was to use the plentiful supply of BBN engineers who had > been debugging Internet crises for several years as the "experts", > observing how they worked, and translating what they did into software > to do tasks "automatically". That started the "Automated Network > Management" (ANM) project as a new research task at BBN. I wrote the > proposal to ARPA and waited for the contract to be signed. But when BBN > reorganized in July 1983 the ANM contract and I went separate ways so I > never got to work on ANM and I don't think there was enough detail in > the proposal to convey the concept onwards to the new project team. > > It strikes me that the real driver of AI today has evolved only partly > from advances in algorithms such as LLMs. I think another major factor > has been the massive drop in computing costs over the last > half-century. AI systems, such as that Morse Code project, were > affordable only by well-funded parts of government in the 1970s, > requiring millions of dollars/pounds/rubles/yen/lira to field a system > that provided minimal capability. Today, the system that cost millions > in the 1970s now costs pennies by comparison. > > Anybody else know more about the history of the use of AI *within* (not > just on top of) The Internet? For example, when there are problems in > today's Internet, are AI techniques and tools used to diagnose and > repair them? What's the History of such things? > > /Jack Haverty > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > From brian.e.carpenter at gmail.com Wed Jun 24 17:36:49 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Thu, 25 Jun 2026 12:36:49 +1200 Subject: [ih] History of AI and Internet In-Reply-To: References: Message-ID: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> On 25-Jun-26 10:22, Sivan via Internet-history wrote: > Dear Jack Haverty, > > Your question "For example, when there are problems in today's Internet, > are AI techniques and tools used to diagnose and repair them? What's the > History of such things?" is immensely interesting. Concerns about A.I. > momentarily set aside, are there initiatives underway to positively use > A.I. tools to "diagnose and repair" problems in the Internet? For example, > using A.I. to scan for malware, bots, phishing and other forms of technical > and non-technical Abuse? Or using A.I. to scan and detect barriers to > network protocols such as vpns? Or even using A.I. to scan and detect > non-human content and other forms of Abuse? I can't imagine that the answer to any of those questions is "No". I wouldn't have said that a year ago, but progress is very fast. Of course, operators will see this as a competitive advantage and may choose not to publicise such AI deployments. But there is a lot of work in progress on agent-to-agent communication. If this is history, it's history in progress. Regards/Ng? mihi Brian Carpenter > > Sivasubramanian Muthusamy > sender > > On Tue, 23 Jun, 2026, 01:20 Jack Haverty via Internet-history, < > internet-history at elists.isoc.org> wrote: > >> With AI now in the news all the time, I've been wondering about the >> history of AI in The Internet, i.e., how AI technology has been used >> within the Internet over the years, rather than just a service which >> people can access over The Internet. Personally I only know of a few >> such uses of AI in a communications role, quite a while ago. But maybe >> others remember more...? >> >> In the mid-1970s, I was involved in an AI project at MIT in Licklider's >> group, sponsored by ARPA. It wasn't really part of the ARPANET but it >> did relate to communications. The project involved using AI techniques >> of that era, then called "expert systems", to have computers decode >> hand-sent Morse radio transmissions. The Intelligence Community was >> apparently quite interested in this problem at that time, and the >> project ran for several years and was deemed a success. >> >> FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 for >> info on that AI project. Personally, I think of that 1990 report as an >> enumeration of "ARPA's Greatest Hits". There are six projects listed >> from the Information Processing Techniques Office of ARPA (IPTO, where >> all the networking and computer projects happened). Another project of >> those six IPTO successes is The ARPANET, described in Chapter 20. The >> Internet was probably still too immature to be included at the time . >> >> Similarly, while I was at BBN in the 1980s there were tools developed to >> visualize activity in the ARPANET, and do tasks involved in network >> design - figuring out where new lines were needed, reconfiguring the >> topology of the ARPANET to address changes in traffic patterns, and >> other such analyses. I'm not sure those projects would be recognized as >> "AI" today, but they were widely used to manage a variety of networks >> such as the ARPANET and DDN. The idea was to use computers to augment >> people skills, as it is in today's AI as well. >> >> Bob Kahn and I had a discussion at some point in late 1982 about AI and >> networks in ARPA's world. Bob was aware of the ARPA-sponsored project >> done at MIT in the mid 1970s which used AI "expert system" techniques to >> decode Morse code. We mused that a similar expert system approach >> could be used to manage networks, especially the emerging Internet which >> was much more complex than the ARPANET and had few tools available for >> operations and management. >> >> The concept was to use the plentiful supply of BBN engineers who had >> been debugging Internet crises for several years as the "experts", >> observing how they worked, and translating what they did into software >> to do tasks "automatically". That started the "Automated Network >> Management" (ANM) project as a new research task at BBN. I wrote the >> proposal to ARPA and waited for the contract to be signed. But when BBN >> reorganized in July 1983 the ANM contract and I went separate ways so I >> never got to work on ANM and I don't think there was enough detail in >> the proposal to convey the concept onwards to the new project team. >> >> It strikes me that the real driver of AI today has evolved only partly >> from advances in algorithms such as LLMs. I think another major factor >> has been the massive drop in computing costs over the last >> half-century. AI systems, such as that Morse Code project, were >> affordable only by well-funded parts of government in the 1970s, >> requiring millions of dollars/pounds/rubles/yen/lira to field a system >> that provided minimal capability. Today, the system that cost millions >> in the 1970s now costs pennies by comparison. >> >> Anybody else know more about the history of the use of AI *within* (not >> just on top of) The Internet? For example, when there are problems in >> today's Internet, are AI techniques and tools used to diagnose and >> repair them? What's the History of such things? >> >> /Jack Haverty >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> From gregskinner0 at icloud.com Wed Jun 24 20:49:39 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Wed, 24 Jun 2026 20:49:39 -0700 Subject: [ih] Fwd: [ainetops] [OPSAREA@IETF126] IRTF/IETF OPS Transfer: Status & Exploring Collaboration Opportunities References: <914962a249574fee8baab3d62d96f33b@huawei.com> Message-ID: The AIOPS related work mentioned in the forwarded message is a link to a GitHub page containing links to current and recent IETF and IRTF AI-related activities. There is even a link to a LISP draft about assigning endpoint-IDs to AI agents. :) --gregbo > Begin forwarded message: > > From: Qin Wu > Subject: [ainetops] Re: [OPSAREA at IETF126] IRTF/IETF OPS Transfer: Status & Exploring Collaboration Opportunities > Date: June 9, 2026 at 7:40:43?PM PDT > To: "Thomas.Graf at swisscom.com" , "nmop at ietf.org" , "nmrg at irtf.org" , "ainetops at ietf.org" > Cc: "mohamed.boucadair at orange.com" , "nmrg-chairs at ietf.org" , "nmop-chairs at ietf.org" , "daniel at olddog.co.uk" > > Thomas: > We see NMRG has invested a lot in Intent, Data, AI in the past and engage with many other SDO/industrial consortium, plan workshops to drive these work. AI are key topics now common to both NMOP and NMRG. Traditional boundary between NMRG and NMOP is that one focuses on protocol and data model while the other focus on architecture and use cases, one focuses on traditional AI while the other looks into the expansion of traditional AI, covering AI for network in the long term evolution. > > Yes, many AIOPS related work (https://github.com/IETF-OPS-AD/AINETOPS) has been distributed in various different places, driven by individual, lack a good coordination, but we do have ainetops mailing list ainetops at ietf.org for AI technologies discussion and coordination, Dan and I recently made update to > https://github.com/IETF-OPS-AD/AINETOPS, to track each AI work and classify them into several categories such as > o Network Operations > o LLM/AI Benchmarking > o Discovery&Naming > o Observability,Intervention and Control > o Security > o Transport > we hope everybody can make good use of this AI coordination mailing list to find home for their work, find people who can work together, or get more input and feedback on their cross area topics from specific area experts in the community. > > Regarding encouraging more exchange between NMRG and NMOP communities, since both Group request 2 sessions in each IETF meeting, I am wondering whether a joint meeting can be planned to make good use of time for both group members. > > Regarding knowledge graph/ontologies, I have been aware that ontologies has gained rapid industry traction becos it connects fragmented data silos and provides Contextual Grounding, Deterministic Reasoning, drive intent-based automation, I know NMOP has drive knowledge graph discussion for a long time with various different challenges. It will be great to see NMRG and NMOP work together in this problem space with moreopen source, implementation, running code. > > In addition, we see little attention being paid to the impact that the introduction of AI may impose on YANG modelling, traditional management protocol, the big challenge introduced by AI is non-deterministic behavior exhibiting by AI or Agentic AI, the question to NMOP WG is are you ready to embrace AI and tackle those challenges? Or these topics should be cooked in various different side meeting first. > > -Qin > From lk at cs.ucla.edu Wed Jun 24 22:12:52 2026 From: lk at cs.ucla.edu (Leonard Kleinrock) Date: Wed, 24 Jun 2026 22:12:52 -0700 (PDT) Subject: [ih] A bit more re John McCarthy In-Reply-To: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> References: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> Message-ID: <9E41BB03-FF53-4324-B27D-34F749B0DE03@cs.ucla.edu> ?Back to the early history involving the brilliance of John McCarthy: For example, it is not widely known that in 1958/59 John McCarthy and Claude Shannon began working on a limited chess playing program. McCarthy engaged a graduate student named Paul Abrahams and they took on the task of generating all legal moves from any given chess position. Shannon hired me to work with him and it was our job to generate strategies for the middle game. In that role, among other strategies, we studied a book by Fred Reinfeld ?1001 brilliant chess sacrifices and combinations? where each page showed a middle game position for which there was a brilliant move or a sequence of moves and the challenge was to find it. We went to the answers in the back of the book and gathered statistics on such data as ?what is the most common first move of brilliant sequence??. (the answer is, unsurprisingly ?check?!). Using such data, we generated a fairly interesting program to play portions of the middle game of chess. Shortly thereafter, Alan Kotok, Elwyn Berlekamp and other MIT students used some of these ideas which led to the MIT chess playing program. As another example, we all remember that McCarthy created LISP (and the same Paul Abrahams was again involved) and McCarthy also introduced some essential ideas of time sharing. Between John McCarthy and Marvin Minsky the field of AI had its early beginnings. Len Sent from my iPhone > On Jun 24, 2026, at 5:36?PM, Brian E Carpenter via Internet-history wrote: > ?On 25-Jun-26 10:22, Sivan via Internet-history wrote: >> Dear Jack Haverty, >> Your question "For example, when there are problems in today's Internet, >> are AI techniques and tools used to diagnose and repair them? What's the >> History of such things?" is immensely interesting. Concerns about A.I. >> momentarily set aside, are there initiatives underway to positively use >> A.I. tools to "diagnose and repair" problems in the Internet? For example, >> using A.I. to scan for malware, bots, phishing and other forms of technical >> and non-technical Abuse? Or using A.I. to scan and detect barriers to >> network protocols such as vpns? Or even using A.I. to scan and detect >> non-human content and other forms of Abuse? > > I can't imagine that the answer to any of those questions is "No". > I wouldn't have said that a year ago, but progress is very fast. > > Of course, operators will see this as a competitive advantage and may > choose not to publicise such AI deployments. But there is a lot of > work in progress on agent-to-agent communication. > > If this is history, it's history in progress. > > Regards/Ng? mihi > Brian Carpenter > >> Sivasubramanian Muthusamy >> sender >>> On Tue, 23 Jun, 2026, 01:20 Jack Haverty via Internet-history, < >>> internet-history at elists.isoc.org> wrote: >>> With AI now in the news all the time, I've been wondering about the >>> history of AI in The Internet, i.e., how AI technology has been used >>> within the Internet over the years, rather than just a service which >>> people can access over The Internet. Personally I only know of a few >>> such uses of AI in a communications role, quite a while ago. But maybe >>> others remember more...? >>> In the mid-1970s, I was involved in an AI project at MIT in Licklider's >>> group, sponsored by ARPA. It wasn't really part of the ARPANET but it >>> did relate to communications. The project involved using AI techniques >>> of that era, then called "expert systems", to have computers decode >>> hand-sent Morse radio transmissions. The Intelligence Community was >>> apparently quite interested in this problem at that time, and the >>> project ran for several years and was deemed a success. >>> FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 for >>> info on that AI project. Personally, I think of that 1990 report as an >>> enumeration of "ARPA's Greatest Hits". There are six projects listed >>> from the Information Processing Techniques Office of ARPA (IPTO, where >>> all the networking and computer projects happened). Another project of >>> those six IPTO successes is The ARPANET, described in Chapter 20. The >>> Internet was probably still too immature to be included at the time . >>> Similarly, while I was at BBN in the 1980s there were tools developed to >>> visualize activity in the ARPANET, and do tasks involved in network >>> design - figuring out where new lines were needed, reconfiguring the >>> topology of the ARPANET to address changes in traffic patterns, and >>> other such analyses. I'm not sure those projects would be recognized as >>> "AI" today, but they were widely used to manage a variety of networks >>> such as the ARPANET and DDN. The idea was to use computers to augment >>> people skills, as it is in today's AI as well. >>> Bob Kahn and I had a discussion at some point in late 1982 about AI and >>> networks in ARPA's world. Bob was aware of the ARPA-sponsored project >>> done at MIT in the mid 1970s which used AI "expert system" techniques to >>> decode Morse code. We mused that a similar expert system approach >>> could be used to manage networks, especially the emerging Internet which >>> was much more complex than the ARPANET and had few tools available for >>> operations and management. >>> The concept was to use the plentiful supply of BBN engineers who had >>> been debugging Internet crises for several years as the "experts", >>> observing how they worked, and translating what they did into software >>> to do tasks "automatically". That started the "Automated Network >>> Management" (ANM) project as a new research task at BBN. I wrote the >>> proposal to ARPA and waited for the contract to be signed. But when BBN >>> reorganized in July 1983 the ANM contract and I went separate ways so I >>> never got to work on ANM and I don't think there was enough detail in >>> the proposal to convey the concept onwards to the new project team. >>> It strikes me that the real driver of AI today has evolved only partly >>> from advances in algorithms such as LLMs. I think another major factor >>> has been the massive drop in computing costs over the last >>> half-century. AI systems, such as that Morse Code project, were >>> affordable only by well-funded parts of government in the 1970s, >>> requiring millions of dollars/pounds/rubles/yen/lira to field a system >>> that provided minimal capability. Today, the system that cost millions >>> in the 1970s now costs pennies by comparison. >>> Anybody else know more about the history of the use of AI *within* (not >>> just on top of) The Internet? For example, when there are problems in >>> today's Internet, are AI techniques and tools used to diagnose and >>> repair them? What's the History of such things? >>> /Jack Haverty >>> -- >>> Internet-history mailing list >>> Internet-history at elists.isoc.org >>> https://elists.isoc.org/mailman/listinfo/internet-history >>> - >>> Unsubscribe: >>> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From gregskinner0 at icloud.com Thu Jun 25 00:33:00 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Thu, 25 Jun 2026 00:33:00 -0700 Subject: [ih] History of AI and Internet In-Reply-To: <1677377271.1720268.1782326721278@mail.yahoo.com> References: <580818707.1307788.1782233544689@mail.yahoo.com> <1677377271.1720268.1782326721278@mail.yahoo.com> Message-ID: <4B5A8925-D722-4AA6-821F-1923D535BEA2@icloud.com> On Jun 24, 2026, at 11:45?AM, Barbara Denny via Internet-history wrote: > > Just to be clear, the second edition was published (2018). I don't know about the draft you read online. I got my copy on amazon. Too bad all the tech bookstores i knew about in the Bay Area closed along time ago. > barbara > I found updated copies (copyright 2018, 2020) of the book on Richard Sutton?s website . The authors made a comment about the rigor of the second edition in the Preface: "As in the first edition, we chose not to produce a rigorous formal treatment of reinforcement learning, or to formulate it in the most general terms. However, our deeper understanding of some topics since the first edition required a bit more mathematics to explain; we have set off the more mathematical parts in shaded boxes that the non- mathematically-inclined may choose to skip." Since some of you are or were professors, I?d be interested in your opinions about the required math background, as the book is designed for a course in reinforcement learning. --gregbo From vint at google.com Thu Jun 25 02:36:18 2026 From: vint at google.com (Vint Cerf) Date: Thu, 25 Jun 2026 05:36:18 -0400 Subject: [ih] A bit more re John McCarthy In-Reply-To: <9E41BB03-FF53-4324-B27D-34F749B0DE03@cs.ucla.edu> References: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> <9E41BB03-FF53-4324-B27D-34F749B0DE03@cs.ucla.edu> Message-ID: Steve Crocker and I spent some time understanding a chess program written by Richard Greenblatt at MIT in the mid/late 1960s. Not sure how that program may have been influenced by earlier work. From Gemini: Alan Kotok, a fellow MIT hacker, wrote an earlier chess program which inspired Greenblatt. The program Richard Greenblatt wrote is famously known as *Mac Hack VI* (or simply *Mac Hack* / *The Greenblatt Chess Program*). Developed at the MIT Artificial Intelligence Laboratory starting in November 1966, it is a milestone in the history of AI and computer chess. Here is a breakdown of what made Mac Hack VI so significant: 1. Historical Milestones - *Human Tournament Play:* In January 1967, Mac Hack VI became the very first computer program to play against humans under regular tournament conditions (competing in the Massachusetts Amateur Championship). - *Official Chess Rating:* It was the first chess engine to be granted an official chess rating by the United States Chess Federation (USCF), initially clocking in around 1200?1400 (the level of a decent hobbyist or high school player). - *First Human Defeat:* In the spring of 1967, it became the first computer program to defeat a human in a sanctioned tournament setting, beating a player rated 1510 USCF. - *The Hubert Dreyfus Match:* Famously, philosopher and prominent AI skeptic Hubert Dreyfus?who had publicly asserted that no computer could ever defeat even a child at chess?was invited to the MIT lab. Mac Hack VI promptly defeated him, a moment that became legendary in early AI lore. 2. Technical Innovations - *The Transposition Table:* Mac Hack VI was the first chess program to implement a transposition table. This optimization caches previously evaluated board positions, ensuring the computer doesn't waste precious processing power re-calculating the same position if arrived at via a different sequence of moves. - *Better Chess Heuristics:* Unlike earlier creators of chess programs who weren't particularly strong players themselves, Richard Greenblatt was an accomplished chess player. He added roughly 50 specialized heuristics (rules of thumb) directly into the code to evaluate positions pragmatically. - *Programming Language:* It was written entirely in *MIDAS* (a macro assembler language) for the DEC PDP-6 computer. It evaluated about 10 positions per second, a fraction of what modern engines do, but used clever "plausible move generation" to narrow down which branches of the game tree to search. 3. Legacy Mac Hack VI went on to be widely distributed after being ported to the PDP-10. It laid the architectural groundwork for many generation-defining chess engines that followed. Interestingly, a decade later in 1977, the program even got to play three games against grandmaster Bobby Fischer (though Fischer predictably won all three). On Thu, Jun 25, 2026 at 1:12?AM Leonard Kleinrock via Internet-history < internet-history at elists.isoc.org> wrote: > Back to the early history involving the brilliance of John McCarthy: For > example, it is not widely known that in 1958/59 John McCarthy and Claude > Shannon began working on a limited chess playing program. McCarthy engaged > a graduate student named Paul Abrahams and they took on the task of > generating all legal moves from any given chess position. Shannon hired me > to work with him and it was our job to generate strategies for the middle > game. In that role, among other strategies, we studied a book by Fred > Reinfeld ?1001 brilliant chess sacrifices and combinations? where each page > showed a middle game position for which there was a brilliant move or a > sequence of moves and the challenge was to find it. We went to the answers > in the back of the book and gathered statistics on such data as ?what is > the most common first move of brilliant sequence??. (the answer is, > unsurprisingly ?check?!). Using such data, we generated a fairly > interesting program to play portions of the middle game of chess. Shortly > thereafter, Alan Kotok, Elwyn Berlekamp and other MIT students used some of > these ideas which led to the MIT chess playing program. > As another example, we all remember that McCarthy created LISP (and the > same Paul Abrahams was again involved) and McCarthy also introduced some > essential ideas of time sharing. > Between John McCarthy and Marvin Minsky the field of AI had its early > beginnings. > Len > Sent from my iPhone > > > > On Jun 24, 2026, at 5:36?PM, Brian E Carpenter via Internet-history < > internet-history at elists.isoc.org> wrote: > > ?On 25-Jun-26 10:22, Sivan via Internet-history wrote: > >> Dear Jack Haverty, > >> Your question "For example, when there are problems in today's Internet, > >> are AI techniques and tools used to diagnose and repair them? What's > the > >> History of such things?" is immensely interesting. Concerns about A.I. > >> momentarily set aside, are there initiatives underway to positively use > >> A.I. tools to "diagnose and repair" problems in the Internet? For > example, > >> using A.I. to scan for malware, bots, phishing and other forms of > technical > >> and non-technical Abuse? Or using A.I. to scan and detect barriers to > >> network protocols such as vpns? Or even using A.I. to scan and detect > >> non-human content and other forms of Abuse? > > > > I can't imagine that the answer to any of those questions is "No". > > I wouldn't have said that a year ago, but progress is very fast. > > > > Of course, operators will see this as a competitive advantage and may > > choose not to publicise such AI deployments. But there is a lot of > > work in progress on agent-to-agent communication. > > > > If this is history, it's history in progress. > > > > Regards/Ng? mihi > > Brian Carpenter > > > >> Sivasubramanian Muthusamy > >> sender > >>> On Tue, 23 Jun, 2026, 01:20 Jack Haverty via Internet-history, < > >>> internet-history at elists.isoc.org> wrote: > >>> With AI now in the news all the time, I've been wondering about the > >>> history of AI in The Internet, i.e., how AI technology has been used > >>> within the Internet over the years, rather than just a service which > >>> people can access over The Internet. Personally I only know of a few > >>> such uses of AI in a communications role, quite a while ago. But maybe > >>> others remember more...? > >>> In the mid-1970s, I was involved in an AI project at MIT in Licklider's > >>> group, sponsored by ARPA. It wasn't really part of the ARPANET but it > >>> did relate to communications. The project involved using AI > techniques > >>> of that era, then called "expert systems", to have computers decode > >>> hand-sent Morse radio transmissions. The Intelligence Community was > >>> apparently quite interested in this problem at that time, and the > >>> project ran for several years and was deemed a success. > >>> FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 > for > >>> info on that AI project. Personally, I think of that 1990 report as an > >>> enumeration of "ARPA's Greatest Hits". There are six projects listed > >>> from the Information Processing Techniques Office of ARPA (IPTO, where > >>> all the networking and computer projects happened). Another project of > >>> those six IPTO successes is The ARPANET, described in Chapter 20. The > >>> Internet was probably still too immature to be included at the time . > >>> Similarly, while I was at BBN in the 1980s there were tools developed > to > >>> visualize activity in the ARPANET, and do tasks involved in network > >>> design - figuring out where new lines were needed, reconfiguring the > >>> topology of the ARPANET to address changes in traffic patterns, and > >>> other such analyses. I'm not sure those projects would be recognized > as > >>> "AI" today, but they were widely used to manage a variety of networks > >>> such as the ARPANET and DDN. The idea was to use computers to augment > >>> people skills, as it is in today's AI as well. > >>> Bob Kahn and I had a discussion at some point in late 1982 about AI and > >>> networks in ARPA's world. Bob was aware of the ARPA-sponsored project > >>> done at MIT in the mid 1970s which used AI "expert system" techniques > to > >>> decode Morse code. We mused that a similar expert system approach > >>> could be used to manage networks, especially the emerging Internet > which > >>> was much more complex than the ARPANET and had few tools available for > >>> operations and management. > >>> The concept was to use the plentiful supply of BBN engineers who had > >>> been debugging Internet crises for several years as the "experts", > >>> observing how they worked, and translating what they did into software > >>> to do tasks "automatically". That started the "Automated Network > >>> Management" (ANM) project as a new research task at BBN. I wrote the > >>> proposal to ARPA and waited for the contract to be signed. But when > BBN > >>> reorganized in July 1983 the ANM contract and I went separate ways so I > >>> never got to work on ANM and I don't think there was enough detail in > >>> the proposal to convey the concept onwards to the new project team. > >>> It strikes me that the real driver of AI today has evolved only partly > >>> from advances in algorithms such as LLMs. I think another major factor > >>> has been the massive drop in computing costs over the last > >>> half-century. AI systems, such as that Morse Code project, were > >>> affordable only by well-funded parts of government in the 1970s, > >>> requiring millions of dollars/pounds/rubles/yen/lira to field a system > >>> that provided minimal capability. Today, the system that cost millions > >>> in the 1970s now costs pennies by comparison. > >>> Anybody else know more about the history of the use of AI *within* (not > >>> just on top of) The Internet? For example, when there are problems in > >>> today's Internet, are AI techniques and tools used to diagnose and > >>> repair them? What's the History of such things? > >>> /Jack Haverty > >>> -- > >>> Internet-history mailing list > >>> Internet-history at elists.isoc.org > >>> https://elists.isoc.org/mailman/listinfo/internet-history > >>> - > >>> Unsubscribe: > >>> > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > -- > > Internet-history mailing list > > Internet-history at elists.isoc.org > > https://elists.isoc.org/mailman/listinfo/internet-history > > - > > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- After July 1, Please send any postal/overnight deliveries to: Vint Cerf 1435 Woodhurst Blvd McLean, VA 22102 Until July 1, send postal/overnight deliveries to: Google, LLC 1900 Reston Metro Plaza, 16th Floor Reston, VA 20190 +1 650-224-2788 From mfidelman at protocoltechnologiesgroup.com Thu Jun 25 04:54:53 2026 From: mfidelman at protocoltechnologiesgroup.com (Miles Fidelman) Date: Thu, 25 Jun 2026 11:54:53 +0000 Subject: [ih] A bit more re John McCarthy In-Reply-To: References: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> <9E41BB03-FF53-4324-B27D-34F749B0DE03@cs.ucla.edu> Message-ID: Re. Hubert Dreyfus: I'm reminded of an MIT AI Lab Note - I believe written by Marvin Minsky - in response to Dreyfus's "Why Computers Can't Play Chess." Either the title, or the first line was "Neither can Dreyfus." Miles ________________________________ From: Internet-history on behalf of Vint Cerf via Internet-history Sent: Thursday, June 25, 2026 5:36 AM To: Leonard Kleinrock Cc: Internet-history Subject: Re: [ih] A bit more re John McCarthy Steve Crocker and I spent some time understanding a chess program written by Richard Greenblatt at MIT in the mid/late 1960s. Not sure how that program may have been influenced by earlier work. From Gemini: Alan Kotok, a fellow MIT hacker, wrote an earlier chess program which inspired Greenblatt. The program Richard Greenblatt wrote is famously known as *Mac Hack VI* (or simply *Mac Hack* / *The Greenblatt Chess Program*). Developed at the MIT Artificial Intelligence Laboratory starting in November 1966, it is a milestone in the history of AI and computer chess. Here is a breakdown of what made Mac Hack VI so significant: 1. Historical Milestones - *Human Tournament Play:* In January 1967, Mac Hack VI became the very first computer program to play against humans under regular tournament conditions (competing in the Massachusetts Amateur Championship). - *Official Chess Rating:* It was the first chess engine to be granted an official chess rating by the United States Chess Federation (USCF), initially clocking in around 1200?1400 (the level of a decent hobbyist or high school player). - *First Human Defeat:* In the spring of 1967, it became the first computer program to defeat a human in a sanctioned tournament setting, beating a player rated 1510 USCF. - *The Hubert Dreyfus Match:* Famously, philosopher and prominent AI skeptic Hubert Dreyfus?who had publicly asserted that no computer could ever defeat even a child at chess?was invited to the MIT lab. Mac Hack VI promptly defeated him, a moment that became legendary in early AI lore. 2. Technical Innovations - *The Transposition Table:* Mac Hack VI was the first chess program to implement a transposition table. This optimization caches previously evaluated board positions, ensuring the computer doesn't waste precious processing power re-calculating the same position if arrived at via a different sequence of moves. - *Better Chess Heuristics:* Unlike earlier creators of chess programs who weren't particularly strong players themselves, Richard Greenblatt was an accomplished chess player. He added roughly 50 specialized heuristics (rules of thumb) directly into the code to evaluate positions pragmatically. - *Programming Language:* It was written entirely in *MIDAS* (a macro assembler language) for the DEC PDP-6 computer. It evaluated about 10 positions per second, a fraction of what modern engines do, but used clever "plausible move generation" to narrow down which branches of the game tree to search. 3. Legacy Mac Hack VI went on to be widely distributed after being ported to the PDP-10. It laid the architectural groundwork for many generation-defining chess engines that followed. Interestingly, a decade later in 1977, the program even got to play three games against grandmaster Bobby Fischer (though Fischer predictably won all three). On Thu, Jun 25, 2026 at 1:12?AM Leonard Kleinrock via Internet-history < internet-history at elists.isoc.org> wrote: > Back to the early history involving the brilliance of John McCarthy: For > example, it is not widely known that in 1958/59 John McCarthy and Claude > Shannon began working on a limited chess playing program. McCarthy engaged > a graduate student named Paul Abrahams and they took on the task of > generating all legal moves from any given chess position. Shannon hired me > to work with him and it was our job to generate strategies for the middle > game. In that role, among other strategies, we studied a book by Fred > Reinfeld ?1001 brilliant chess sacrifices and combinations? where each page > showed a middle game position for which there was a brilliant move or a > sequence of moves and the challenge was to find it. We went to the answers > in the back of the book and gathered statistics on such data as ?what is > the most common first move of brilliant sequence??. (the answer is, > unsurprisingly ?check?!). Using such data, we generated a fairly > interesting program to play portions of the middle game of chess. Shortly > thereafter, Alan Kotok, Elwyn Berlekamp and other MIT students used some of > these ideas which led to the MIT chess playing program. > As another example, we all remember that McCarthy created LISP (and the > same Paul Abrahams was again involved) and McCarthy also introduced some > essential ideas of time sharing. > Between John McCarthy and Marvin Minsky the field of AI had its early > beginnings. > Len > Sent from my iPhone > > > > On Jun 24, 2026, at 5:36?PM, Brian E Carpenter via Internet-history < > internet-history at elists.isoc.org> wrote: > > ?On 25-Jun-26 10:22, Sivan via Internet-history wrote: > >> Dear Jack Haverty, > >> Your question "For example, when there are problems in today's Internet, > >> are AI techniques and tools used to diagnose and repair them? What's > the > >> History of such things?" is immensely interesting. Concerns about A.I. > >> momentarily set aside, are there initiatives underway to positively use > >> A.I. tools to "diagnose and repair" problems in the Internet? For > example, > >> using A.I. to scan for malware, bots, phishing and other forms of > technical > >> and non-technical Abuse? Or using A.I. to scan and detect barriers to > >> network protocols such as vpns? Or even using A.I. to scan and detect > >> non-human content and other forms of Abuse? > > > > I can't imagine that the answer to any of those questions is "No". > > I wouldn't have said that a year ago, but progress is very fast. > > > > Of course, operators will see this as a competitive advantage and may > > choose not to publicise such AI deployments. But there is a lot of > > work in progress on agent-to-agent communication. > > > > If this is history, it's history in progress. > > > > Regards/Ng? mihi > > Brian Carpenter > > > >> Sivasubramanian Muthusamy > >> sender > >>> On Tue, 23 Jun, 2026, 01:20 Jack Haverty via Internet-history, < > >>> internet-history at elists.isoc.org> wrote: > >>> With AI now in the news all the time, I've been wondering about the > >>> history of AI in The Internet, i.e., how AI technology has been used > >>> within the Internet over the years, rather than just a service which > >>> people can access over The Internet. Personally I only know of a few > >>> such uses of AI in a communications role, quite a while ago. But maybe > >>> others remember more...? > >>> In the mid-1970s, I was involved in an AI project at MIT in Licklider's > >>> group, sponsored by ARPA. It wasn't really part of the ARPANET but it > >>> did relate to communications. The project involved using AI > techniques > >>> of that era, then called "expert systems", to have computers decode > >>> hand-sent Morse radio transmissions. The Intelligence Community was > >>> apparently quite interested in this problem at that time, and the > >>> project ran for several years and was deemed a success. > >>> FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 > for > >>> info on that AI project. Personally, I think of that 1990 report as an > >>> enumeration of "ARPA's Greatest Hits". There are six projects listed > >>> from the Information Processing Techniques Office of ARPA (IPTO, where > >>> all the networking and computer projects happened). Another project of > >>> those six IPTO successes is The ARPANET, described in Chapter 20. The > >>> Internet was probably still too immature to be included at the time . > >>> Similarly, while I was at BBN in the 1980s there were tools developed > to > >>> visualize activity in the ARPANET, and do tasks involved in network > >>> design - figuring out where new lines were needed, reconfiguring the > >>> topology of the ARPANET to address changes in traffic patterns, and > >>> other such analyses. I'm not sure those projects would be recognized > as > >>> "AI" today, but they were widely used to manage a variety of networks > >>> such as the ARPANET and DDN. The idea was to use computers to augment > >>> people skills, as it is in today's AI as well. > >>> Bob Kahn and I had a discussion at some point in late 1982 about AI and > >>> networks in ARPA's world. Bob was aware of the ARPA-sponsored project > >>> done at MIT in the mid 1970s which used AI "expert system" techniques > to > >>> decode Morse code. We mused that a similar expert system approach > >>> could be used to manage networks, especially the emerging Internet > which > >>> was much more complex than the ARPANET and had few tools available for > >>> operations and management. > >>> The concept was to use the plentiful supply of BBN engineers who had > >>> been debugging Internet crises for several years as the "experts", > >>> observing how they worked, and translating what they did into software > >>> to do tasks "automatically". That started the "Automated Network > >>> Management" (ANM) project as a new research task at BBN. I wrote the > >>> proposal to ARPA and waited for the contract to be signed. But when > BBN > >>> reorganized in July 1983 the ANM contract and I went separate ways so I > >>> never got to work on ANM and I don't think there was enough detail in > >>> the proposal to convey the concept onwards to the new project team. > >>> It strikes me that the real driver of AI today has evolved only partly > >>> from advances in algorithms such as LLMs. I think another major factor > >>> has been the massive drop in computing costs over the last > >>> half-century. AI systems, such as that Morse Code project, were > >>> affordable only by well-funded parts of government in the 1970s, > >>> requiring millions of dollars/pounds/rubles/yen/lira to field a system > >>> that provided minimal capability. Today, the system that cost millions > >>> in the 1970s now costs pennies by comparison. > >>> Anybody else know more about the history of the use of AI *within* (not > >>> just on top of) The Internet? For example, when there are problems in > >>> today's Internet, are AI techniques and tools used to diagnose and > >>> repair them? What's the History of such things? > >>> /Jack Haverty > >>> -- > >>> Internet-history mailing list > >>> Internet-history at elists.isoc.org > >>> https://elists.isoc.org/mailman/listinfo/internet-history > >>> - > >>> Unsubscribe: > >>> > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > > -- > > Internet-history mailing list > > Internet-history at elists.isoc.org > > https://elists.isoc.org/mailman/listinfo/internet-history > > - > > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- > Internet-history mailing list > Internet-history at elists.isoc.org > https://elists.isoc.org/mailman/listinfo/internet-history > - > Unsubscribe: > https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history > -- After July 1, Please send any postal/overnight deliveries to: Vint Cerf 1435 Woodhurst Blvd McLean, VA 22102 Until July 1, send postal/overnight deliveries to: Google, LLC 1900 Reston Metro Plaza, 16th Floor Reston, VA 20190 +1 650-224-2788 -- Internet-history mailing list Internet-history at elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history - Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history From 6.internet at gmail.com Thu Jun 25 06:28:20 2026 From: 6.internet at gmail.com (Sivan) Date: Thu, 25 Jun 2026 18:58:20 +0530 Subject: [ih] History of AI and Internet In-Reply-To: References: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> Message-ID: Thank you Brian and Andrew. Sivasubramanian Muthusamy sender On Thu, 25 Jun, 2026, 18:18 Andrew G. Malis, wrote: > Brian et al, > > You're right that progress is very fast. I just saw this article yesterday about > KDDI and Samsung using a mobile network digital twin to test AI-based > network optimization and autonomous operation without disrupting the > production network. > > > https://www.sdxcentral.com/news/nvidia-and-samsung-to-power-6g-ready-digital-twin-for-japan-telecom-giant/ > > Cheers, > Andy > > On Wed, Jun 24, 2026 at 8:37?PM Brian E Carpenter via Internet-history < > internet-history at elists.isoc.org> wrote: > >> On 25-Jun-26 10:22, Sivan via Internet-history wrote: >> > Dear Jack Haverty, >> > >> > Your question "For example, when there are problems in today's Internet, >> > are AI techniques and tools used to diagnose and repair them? What's >> the >> > History of such things?" is immensely interesting. Concerns about A.I. >> > momentarily set aside, are there initiatives underway to positively use >> > A.I. tools to "diagnose and repair" problems in the Internet? For >> example, >> > using A.I. to scan for malware, bots, phishing and other forms of >> technical >> > and non-technical Abuse? Or using A.I. to scan and detect barriers to >> > network protocols such as vpns? Or even using A.I. to scan and detect >> > non-human content and other forms of Abuse? >> >> I can't imagine that the answer to any of those questions is "No". >> I wouldn't have said that a year ago, but progress is very fast. >> >> Of course, operators will see this as a competitive advantage and may >> choose not to publicise such AI deployments. But there is a lot of >> work in progress on agent-to-agent communication. >> >> If this is history, it's history in progress. >> >> Regards/Ng? mihi >> Brian Carpenter >> >> >> From jack at 3kitty.org Thu Jun 25 11:42:08 2026 From: jack at 3kitty.org (Jack Haverty) Date: Thu, 25 Jun 2026 11:42:08 -0700 Subject: [ih] A bit more re John McCarthy In-Reply-To: References: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> <9E41BB03-FF53-4324-B27D-34F749B0DE03@cs.ucla.edu> Message-ID: <89c69ac0-0de7-48ad-939a-073f4a8ad733@3kitty.org> One additional historical tidbit that Gemini missed... The PDP-10 instruction set included ROTC - ROTateCombined - which treated two registers as a single memory location and ROTated bits between the two.? IIRC bits ROTated off the right end of one register were moved into the right side of the adjacent register. Same migration for the other sides, so the 2 registers behaved as one circular register twice as wide.? This was used in Greenblatt's Chess program in one of the "inner loops" that dictated performance.? ?I suspect ROTC might have been used in the "transposition table" manipulations. Greenblatt at one point decided it would also be useful for Chess to have a different instruction available - where the bits rotated out the right end went into the LEFT side of the adjacent register. So one day shortly thereafter the AI's PDP-10 suddenly had a new instruction, which rotated 2 registers "the other way".? I don't remember what mnemonic they used.? I was in Lick's group, and our PDP-10 didn't get that modification so I never used that instruction. This was all possible because the PDP-10s used "flip chips" containing logic gates, all interconnected by a massive "bed of nails" backplane.? ?The CPU was contained in a very large standard DEC cabinet.? So it was straightforward to change how the CPU worked.? If you knew what you were doing you could even add new instructions.? ?DEC Field Service technicians hated coming to fix MIT machines.? But loved to learn more about how the machines actually worked. With today's technology, all that design is now embedded in the silicon wafers, so it would be much much more difficult to do such things. /Jack Haverty On 6/25/26 02:36, Vint Cerf via Internet-history wrote: > Steve Crocker and I spent some time understanding a chess program written > by Richard Greenblatt at MIT in the mid/late 1960s. Not sure how that > program may have been influenced by earlier work. From Gemini: > > Alan Kotok, a fellow MIT hacker, wrote an earlier chess program which > inspired Greenblatt. > > The program Richard Greenblatt wrote is famously known as *Mac Hack VI* (or > simply *Mac Hack* / *The Greenblatt Chess Program*). Developed at the MIT > Artificial Intelligence Laboratory starting in November 1966, it is a > milestone in the history of AI and computer chess. > > Here is a breakdown of what made Mac Hack VI so significant: > 1. Historical Milestones > > - > > *Human Tournament Play:* In January 1967, Mac Hack VI became the very > first computer program to play against humans under regular tournament > conditions (competing in the Massachusetts Amateur Championship). > - > > *Official Chess Rating:* It was the first chess engine to be granted an > official chess rating by the United States Chess Federation (USCF), > initially clocking in around 1200?1400 (the level of a decent hobbyist or > high school player). > - > > *First Human Defeat:* In the spring of 1967, it became the first > computer program to defeat a human in a sanctioned tournament setting, > beating a player rated 1510 USCF. > - > > *The Hubert Dreyfus Match:* Famously, philosopher and prominent AI > skeptic Hubert Dreyfus?who had publicly asserted that no computer could > ever defeat even a child at chess?was invited to the MIT lab. Mac Hack > VI promptly defeated him, a moment that became legendary in early AI lore. > > 2. Technical Innovations > > - > > *The Transposition Table:* Mac Hack VI was the first chess program to > implement a transposition table. This optimization caches previously > evaluated board positions, ensuring the computer doesn't waste precious > processing power re-calculating the same position if arrived at via a > different sequence of moves. > - > > *Better Chess Heuristics:* Unlike earlier creators of chess programs who > weren't particularly strong players themselves, Richard Greenblatt was an > accomplished chess player. He added roughly 50 specialized heuristics > (rules of thumb) directly into the code to evaluate positions pragmatically. > - > > *Programming Language:* It was written entirely in *MIDAS* (a macro > assembler language) for the DEC PDP-6 computer. It evaluated about 10 > positions per second, a fraction of what modern engines do, but used clever > "plausible move generation" to narrow down which branches of the game tree > to search. > > 3. Legacy > > Mac Hack VI went on to be widely distributed after being ported to the > PDP-10. It laid the architectural groundwork for many generation-defining > chess engines that followed. Interestingly, a decade later in 1977, the > program even got to play three games against grandmaster Bobby Fischer > (though Fischer predictably won all three). > > > > On Thu, Jun 25, 2026 at 1:12?AM Leonard Kleinrock via Internet-history < > internet-history at elists.isoc.org> wrote: > >> Back to the early history involving the brilliance of John McCarthy: For >> example, it is not widely known that in 1958/59 John McCarthy and Claude >> Shannon began working on a limited chess playing program. McCarthy engaged >> a graduate student named Paul Abrahams and they took on the task of >> generating all legal moves from any given chess position. Shannon hired me >> to work with him and it was our job to generate strategies for the middle >> game. In that role, among other strategies, we studied a book by Fred >> Reinfeld ?1001 brilliant chess sacrifices and combinations? where each page >> showed a middle game position for which there was a brilliant move or a >> sequence of moves and the challenge was to find it. We went to the answers >> in the back of the book and gathered statistics on such data as ?what is >> the most common first move of brilliant sequence??. (the answer is, >> unsurprisingly ?check?!). Using such data, we generated a fairly >> interesting program to play portions of the middle game of chess. Shortly >> thereafter, Alan Kotok, Elwyn Berlekamp and other MIT students used some of >> these ideas which led to the MIT chess playing program. >> As another example, we all remember that McCarthy created LISP (and the >> same Paul Abrahams was again involved) and McCarthy also introduced some >> essential ideas of time sharing. >> Between John McCarthy and Marvin Minsky the field of AI had its early >> beginnings. >> Len >> Sent from my iPhone >> >> >>> On Jun 24, 2026, at 5:36?PM, Brian E Carpenter via Internet-history < >> internet-history at elists.isoc.org> wrote: >>> ?On 25-Jun-26 10:22, Sivan via Internet-history wrote: >>>> Dear Jack Haverty, >>>> Your question "For example, when there are problems in today's Internet, >>>> are AI techniques and tools used to diagnose and repair them? What's >> the >>>> History of such things?" is immensely interesting. Concerns about A.I. >>>> momentarily set aside, are there initiatives underway to positively use >>>> A.I. tools to "diagnose and repair" problems in the Internet? For >> example, >>>> using A.I. to scan for malware, bots, phishing and other forms of >> technical >>>> and non-technical Abuse? Or using A.I. to scan and detect barriers to >>>> network protocols such as vpns? Or even using A.I. to scan and detect >>>> non-human content and other forms of Abuse? >>> I can't imagine that the answer to any of those questions is "No". >>> I wouldn't have said that a year ago, but progress is very fast. >>> >>> Of course, operators will see this as a competitive advantage and may >>> choose not to publicise such AI deployments. But there is a lot of >>> work in progress on agent-to-agent communication. >>> >>> If this is history, it's history in progress. >>> >>> Regards/Ng? mihi >>> Brian Carpenter >>> >>>> Sivasubramanian Muthusamy >>>> sender >>>>> On Tue, 23 Jun, 2026, 01:20 Jack Haverty via Internet-history, < >>>>> internet-history at elists.isoc.org> wrote: >>>>> With AI now in the news all the time, I've been wondering about the >>>>> history of AI in The Internet, i.e., how AI technology has been used >>>>> within the Internet over the years, rather than just a service which >>>>> people can access over The Internet. Personally I only know of a few >>>>> such uses of AI in a communications role, quite a while ago. But maybe >>>>> others remember more...? >>>>> In the mid-1970s, I was involved in an AI project at MIT in Licklider's >>>>> group, sponsored by ARPA. It wasn't really part of the ARPANET but it >>>>> did relate to communications. The project involved using AI >> techniques >>>>> of that era, then called "expert systems", to have computers decode >>>>> hand-sent Morse radio transmissions. The Intelligence Community was >>>>> apparently quite interested in this problem at that time, and the >>>>> project ran for several years and was deemed a success. >>>>> FYI, see chapter 22 of https://apps.dtic.mil/sti/citations/ADA239925 >> for >>>>> info on that AI project. Personally, I think of that 1990 report as an >>>>> enumeration of "ARPA's Greatest Hits". There are six projects listed >>>>> from the Information Processing Techniques Office of ARPA (IPTO, where >>>>> all the networking and computer projects happened). Another project of >>>>> those six IPTO successes is The ARPANET, described in Chapter 20. The >>>>> Internet was probably still too immature to be included at the time . >>>>> Similarly, while I was at BBN in the 1980s there were tools developed >> to >>>>> visualize activity in the ARPANET, and do tasks involved in network >>>>> design - figuring out where new lines were needed, reconfiguring the >>>>> topology of the ARPANET to address changes in traffic patterns, and >>>>> other such analyses. I'm not sure those projects would be recognized >> as >>>>> "AI" today, but they were widely used to manage a variety of networks >>>>> such as the ARPANET and DDN. The idea was to use computers to augment >>>>> people skills, as it is in today's AI as well. >>>>> Bob Kahn and I had a discussion at some point in late 1982 about AI and >>>>> networks in ARPA's world. Bob was aware of the ARPA-sponsored project >>>>> done at MIT in the mid 1970s which used AI "expert system" techniques >> to >>>>> decode Morse code. We mused that a similar expert system approach >>>>> could be used to manage networks, especially the emerging Internet >> which >>>>> was much more complex than the ARPANET and had few tools available for >>>>> operations and management. >>>>> The concept was to use the plentiful supply of BBN engineers who had >>>>> been debugging Internet crises for several years as the "experts", >>>>> observing how they worked, and translating what they did into software >>>>> to do tasks "automatically". That started the "Automated Network >>>>> Management" (ANM) project as a new research task at BBN. I wrote the >>>>> proposal to ARPA and waited for the contract to be signed. But when >> BBN >>>>> reorganized in July 1983 the ANM contract and I went separate ways so I >>>>> never got to work on ANM and I don't think there was enough detail in >>>>> the proposal to convey the concept onwards to the new project team. >>>>> It strikes me that the real driver of AI today has evolved only partly >>>>> from advances in algorithms such as LLMs. I think another major factor >>>>> has been the massive drop in computing costs over the last >>>>> half-century. AI systems, such as that Morse Code project, were >>>>> affordable only by well-funded parts of government in the 1970s, >>>>> requiring millions of dollars/pounds/rubles/yen/lira to field a system >>>>> that provided minimal capability. Today, the system that cost millions >>>>> in the 1970s now costs pennies by comparison. >>>>> Anybody else know more about the history of the use of AI *within* (not >>>>> just on top of) The Internet? For example, when there are problems in >>>>> today's Internet, are AI techniques and tools used to diagnose and >>>>> repair them? What's the History of such things? >>>>> /Jack Haverty >>>>> -- >>>>> Internet-history mailing list >>>>> Internet-history at elists.isoc.org >>>>> https://elists.isoc.org/mailman/listinfo/internet-history >>>>> - >>>>> Unsubscribe: >>>>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >>> -- >>> Internet-history mailing list >>> Internet-history at elists.isoc.org >>> https://elists.isoc.org/mailman/listinfo/internet-history >>> - >>> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> -- >> Internet-history mailing list >> Internet-history at elists.isoc.org >> https://elists.isoc.org/mailman/listinfo/internet-history >> - >> Unsubscribe: >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history >> > -------------- next part -------------- A non-text attachment was scrubbed... Name: OpenPGP_signature.asc Type: application/pgp-signature Size: 665 bytes Desc: OpenPGP digital signature URL: From brian.e.carpenter at gmail.com Thu Jun 25 15:27:27 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Fri, 26 Jun 2026 10:27:27 +1200 Subject: [ih] History of AI and Internet In-Reply-To: <4B5A8925-D722-4AA6-821F-1923D535BEA2@icloud.com> References: <580818707.1307788.1782233544689@mail.yahoo.com> <1677377271.1720268.1782326721278@mail.yahoo.com> <4B5A8925-D722-4AA6-821F-1923D535BEA2@icloud.com> Message-ID: Greg, > Since some of you are or were professors, I?d be interested in your opinions about the required math background, as the book is designed for a course in reinforcement learning. In my opinion, the book is a masterpiece, but it needs fluency in differential calculus and matrix algebra (which I lost many years ago). I think it would be quite hard to read cover-to-cover while skipping the maths. Also, if you're interested in applicability to large language models, it won't help you. A lot has happened since 2020. Ananthaswamy's book that I mentioned recently relies on similar maths, but you can largely skip it without losing the thread. Regards/Ng? mihi Brian Carpenter On 25-Jun-26 19:33, Greg Skinner via Internet-history wrote: > On Jun 24, 2026, at 11:45?AM, Barbara Denny via Internet-history wrote: >> >> Just to be clear, the second edition was published (2018). I don't know about the draft you read online. I got my copy on amazon. Too bad all the tech bookstores i knew about in the Bay Area closed along time ago. >> barbara >> > > I found updated copies (copyright 2018, 2020) of the book on Richard Sutton?s website . The authors made a comment about the rigor of the second edition in the Preface: > > "As in the first edition, we chose not to produce a rigorous formal treatment of > reinforcement learning, or to formulate it in the most general terms. However, our deeper > understanding of some topics since the first edition required a bit more mathematics > to explain; we have set off the more mathematical parts in shaded boxes that the non- > mathematically-inclined may choose to skip." > > Since some of you are or were professors, I?d be interested in your opinions about the required math background, as the book is designed for a course in reinforcement learning. > > --gregbo > From jack at 3kitty.org Thu Jun 25 15:27:30 2026 From: jack at 3kitty.org (Jack Haverty) Date: Thu, 25 Jun 2026 15:27:30 -0700 Subject: [ih] History of AI and Internet In-Reply-To: References: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> Message-ID: <721f4883-ee57-46a5-bffa-ccdfba6450fa@3kitty.org> Thanks, Arun. I agree that network operators have always used available tools, whatever they might be called.? One could even argue that the ARPANET, 50+ years ago, used AI -- the routing mechanisms and algorithms could alter traffic paths much faster than a human with a patch panel even then, and did so without human intervention required at all. In the early days of The Internet, circa late 1970s, connectivity between Europe and the US was provided by SATNET, which used a channel on the Intelsat IV-A satellite for transatlantic paths.? At the time, the bulk of other users' traffic using Intelsat IV-A was video feeds and telephone calls.? SATNET was likely one of the first to send data over the satellite, with computers at each end (called SIMPs for Satellite IMPs). The SIMPs performed constant checks on the error behavior of the satellite channel, just as the IMPs in the ARPANET measured circuit behavior.? Such checks were performed constantly and consistently, not just when a (human) technician was investigating a problem.? The BBN NOC monitored SATNET as well as ARPANET, so the NOC operators would be informed with a loud bell from the TTY in the NOC when a circuit was degrading either in ARPANET or SATNET. ARPANET circuit degradations were usually referred to ATT Long Lines, where most circuits were provisioned.? SATNET circuit degradations were reported to the Intelsat IV-A NOC. At first, the Intelsat NOC simply dismissed such reports.? It was difficult for them to believe that some unknown company (BBN) in Massachusetts could know anything about the performance of their satellite channel between ground stations in West Virginia and Goonhilly Downs in Europe.? ?Users of video or telephony didn't notice a few glitches in their screens or audio, until it got bad enough.? But the SIMPs did notice right away.? It didn't take long for the Intelsat NOC to start paying attention to such reports, which often preceded some kind of major failure in their satellite service.? The Internet became an "early warning" service for video and telephony services through the satellite. I would characterize such early mechanisms as "expert systems" variety of AI.? They all did what humans would do but, unlike humans, they could do the work much faster and work 24 hours a day. Along with the memory of the ARPA project to use expert systems techniques in a Morse Code communications environment, these experiences motivated the pursuit of the "Automated Network Management" project, to apply such expert systems techniques to the operation and management of networks.? ?I believed it was possible even in 1982.? ?Still do. Fast forward to modern times.... I've been thinking about network problems for quite a while now, but from an end-users' perspective.? ?End-users have lots of "network problems", and as a local tech nerd (surrounded in my neighborhood by non-techies) I get such questions.? I've been experimenting with various AIs to learn what they can do, so I just asked one of them a typical end-user Networking question and asked it how to fix it. The answer I got was unsurprising.? My summary of the AI -- "There is no one to call and report your problem.? It could be a number of things.? You're on your own to figure it out."? If anyone's curious, here's the conversation I just had: https://claude.ai/share/468f2576-bd1b-4dbb-9a48-941a94051b2b Surely The Internet and AI can do better...?? ?I still think some Expert System approach might be appropriate, even if it is an ancient idea. --------------------- One possible cause of the 50 year timeline for AI to be more broadly used in Networking is what I term "Technology Silos". During the 1970s/1980s, there was a lot of research on both AI and Networking.? ?At MIT both were quite active, but in separate groups.? The MIT AI Lab was the locus of AI research.? Licklider's group and a few others (such as Multics) did lots of Network research projects.? The groups were friendly and collaborated on many things, such as our underlying operating system (ITS). But there was little crosstalk in core research issues of either Networking or AI.? AI was interesting to Networkers (like me), but not our Mission.? Networking was interesting to the AI groups, but not their Mission either.? ?It wasn't until the Morse Project appeared in our Networking group that I really had a reason to understand what "Expert Systems" was all about.? ?We successfully used "Expert Systems" to create a computer system.? The government report concluded 'Using available AI techniques, a successful automatic "Morse Code reader" was developed by the MIT group and picked up quickly by NSA.'? ?I've always thought that the goal of Research should include getting the results into actual use "in the field."? ?The introduction of the "Technology Silos" of Morse Code (a form of Networking) and AI was the enabler.? ? Having a distinct goal of "Understanding Morse Code" focussed our research on the actual problem the customer wanted to be solved. I encountered another "Technology Silo" in 1990, when I joined Oracle in Silicon Valley.? The group I joined developed software for use on all types of networks.? But it also had responsibility for operating the corporate "intranet", which had cisco routers deployed in more than 100 countries.? As you might expect, there were lots of "network problems" and our operators struggled to diagnose and fix them.? The tools available from the Networking industry weren't much help. One afternoon, two of us sat down at a console to help diagnose some problem.? I knew a lot about Networking from my years at BBN and Internet experience.? My colleague, like everyone else at my new employer, knew a lot about databases. We couldn't modify any of the routers of course, so changing any of the existing code or protocols was impractical.? But I mentioned SNMP as a mechanism we had used in The Internet to gather data about what was happening in the network.? IIRC we were sitting at a Sun Workstation (1990 vintage), and it had tools such as "snmpget". ?The 1990 routers also "supported SNMP", whatever that meant at the time.? We even found a few host computers (such as the Sun) which had some implementations of SNMP, critical for figuring out what TCPs were doing. In a few hours, literally, we had written some small shell scripts that continuously probed routers and hosts -- anything that would respond to SNMP interactions.? All the data was entered into a database.? IIRC every tidbit of SNMP data from the field was simply entered as a row in a table. Once the data was in a database table, it looked just like all other sorts of data that companies used to manage their business operations - inventory, shipping, customer records, etc. The ARPANET, and The Internet, were created as ARPA projects, which funded much of the early Networking implementations.? ?Most Networkers may not know that database technology also had its basis in government-funded projects.? ?For example, Oracle's genesis (long before I got there) involved funding from the Intelligence parts of the US government (as I suspect the Morse Code project was). So by 1990 there had been more than a decade of databases being evolved and used to manage all sorts of business data.? Networking data gathered through SNMP was just another form of business data. ?Database people have long experience in managing all sorts of business data. I looked into the NOC at Oracle a week or so later, and was wowed at what I saw.? Database people had clearly been there, and all of Oracle's technology was of course available for them to use.? ?I saw a screen with a scrolling display showing network performance data, constantly being updated as more data arrived from SNMP.? It looked like the various medical screens you see in hospitals monitoring lifesigns or in Investment Banks watching stock prices and trends. ? ?You could also easily bring up a screen showing current performance and also display similar graphs of past performance to see how things were different when there were no problems. None of this was what I would call AI, even 35 years ago.? But it was useful.? The key enabling action was to link together the two longstanding "Technology Silos", of Networking and Databases, which had both existed for a long time but apparently not been coupled together. AI has now advanced to include the ability to interact with natural human languages.? There's no longer a need to learn some programming language such as LISP or whatever.? That makes AI much more accessible to people outside of the AI community.? Computing costs have also come way down. So perhaps "Technology Silos" of AI and Networking have been a longstanding obstacle which is now disappearing. /Jack Haverty On 6/25/26 09:19, Arun Welch wrote: > >> On Jun 24, 2026, at 7:36?PM, Brian E Carpenter via Internet-history wrote: >> >> On 25-Jun-26 10:22, Sivan via Internet-history wrote: >>> Dear Jack Haverty, >>> Your question "For example, when there are problems in today's Internet, >>> are AI techniques and tools used to diagnose and repair them? What's the >>> History of such things?" is immensely interesting. Concerns about A.I. >>> momentarily set aside, are there initiatives underway to positively use >>> A.I. tools to "diagnose and repair" problems in the Internet? For example, >>> using A.I. to scan for malware, bots, phishing and other forms of technical >>> and non-technical Abuse? Or using A.I. to scan and detect barriers to >>> network protocols such as vpns? Or even using A.I. to scan and detect >>> non-human content and other forms of Abuse? >> I can't imagine that the answer to any of those questions is "No". >> I wouldn't have said that a year ago, but progress is very fast. > Large operators have always used every technology available, and this has included AI (for whatever the current interpretation of that term is) for a long time. You simply cannot run at scale using only human support. Reducing noise and swivel-chair in the NOC, rapid issue resolution, stability, etc become harder as the network gets larger. > > >> Of course, operators will see this as a competitive advantage and may >> choose not to publicise such AI deployments. But there is a lot of >> work in progress on agent-to-agent communication. > > In our case, we see talking about it as a competitive advantage, and our VP in charge of all this regularly speaks about what we?re doing at conferences, on LinkedIn, etc. As one of the people in our company who spends a lot of time talking to customers about what we?re doing, I can tell you that they absolutely love to hear about this. > > To provide some concrete examples on how we?re using AI today: > > a) Incident correlation: If the circuit is down, you don?t need to troubleshoot BGP. You?ll get alerts for both, and choosing which to safely ignore is important. > b) Predictive analytics: We?re gathering metrics on lots of things, and can predict that a laser diode is going to fail because we?ve seen the light levels drop with this pattern before. We can schedule an outage and replace the hardware in a controlled manner rather than an unplanned outage. > c) Chatbots for our service and NOC operators: ?What?s the status of site/circuit/incident X??. This uses agents to gather information across multiple systems to present a single cohesive answer, reducing swivel chair. In addition to our internal monitoring systems we?re also querying weather, 811 (call before you dig) registries, power systems, etc. to get a comprehensive view. For example, if there?s a regional power outage then you probably know why a site is down. Some of this is also customer-facing, so that a customer can send an email asking ?what?s the status of my outage at 100 Main Street? resulting in a response that lays out all the information known about the issue. > d) Analyzing traffic patterns to shut down botnets. > e) Writing automation workflows using AI tools. Spec-driven development has some interesting implications here. > f) Using AI coding tools to develop other sorts of things. This is particularly true for doing dashboards and analysis. > > The list goes on, this is not everything we?re doing, AI in many forms is used for all sorts of things. > > I?m not speaking in any official capacity here, obviously. > > ?arun -------------- next part -------------- A non-text attachment was scrubbed... Name: OpenPGP_signature.asc Type: application/pgp-signature Size: 665 bytes Desc: OpenPGP digital signature URL: From lars at nocrew.org Thu Jun 25 22:19:56 2026 From: lars at nocrew.org (Lars Brinkhoff) Date: Fri, 26 Jun 2026 05:19:56 +0000 Subject: [ih] A bit more re John McCarthy In-Reply-To: <89c69ac0-0de7-48ad-939a-073f4a8ad733@3kitty.org> (Jack Haverty via Internet-history's message of "Thu, 25 Jun 2026 11:42:08 -0700") References: <21aed31d-893c-402c-ae6f-d636dc7c481e@gmail.com> <9E41BB03-FF53-4324-B27D-34F749B0DE03@cs.ucla.edu> <89c69ac0-0de7-48ad-939a-073f4a8ad733@3kitty.org> Message-ID: <7wwlvlx4hv.fsf@junk.nocrew.org> Jack Haverty wrote: > The PDP-10 instruction set included ROTC - ROTateCombined - which > treated two registers as a single memory location and ROTated bits > between the two. IIRC bits ROTated off the right end of one register > were moved into the right side of the adjacent register. [...] > Greenblatt at one point decided it would also be useful for Chess to > have a different instruction available - where the bits rotated out > the right end went into the LEFT side of the adjacent register. This calls for an infodump! Positive shifts rotate to the left, negative to the right. (On the PDP-10, "left" and "right" are firmly associated with "more significant" and "less significant".) If A and B are adjacent accumulators (registers), rotating one step looks like this. Will only look good with a monospace font. ROTC A,1 ---------------------------------------------------- | -------------------- -------------------- | -| AAAAAAAAAAAAAAAAAB | <- | BBBBBBBBBBBBBBBBBA | <- -------------------- -------------------- The new instruction was called CIRC, and one step looks like this. CIRC A,1 ---------------------------- | -------------------- | -------------------- -| AAAAAAAAAAAAAAAAAB | <- -> | ABBBBBBBBBBBBBBBBB | - -------------------- | -------------------- | ----------------------------- Another useful feature hacked into the MIT machines was a software controlled instruction single step. (Single stepping could be done from the front panel, but that's not useful for debuggers under timesharing.) The above additions were also made to the KL10 microcode running on MIT-MC, and the micocode on the four replacement KS10 machines later. Another much more obscure addition to the instruction set was a PC-relative addressing mode. It was only available on the AI lab PDP-10, but I have not seen any code that uses it. From gregskinner0 at icloud.com Fri Jun 26 01:32:21 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Fri, 26 Jun 2026 01:32:21 -0700 Subject: [ih] Fwd: History of AI and Internet References: <755299404.2218793.1782443866595@mail.yahoo.com> Message-ID: Forwarded for Barbara > On Thursday, June 25, 2026 at 08:08:57 PM PDT, Barbara Denny wrote: > > > Hi > > I have heard things contrary to what i think you said regarding Reinforcement Learning and its applicability to today's LLMs > > Here is what Gemini told me but i do know sometimes it makes big mistakes. I still find it interesting to see what it put together. Too far out of my experience to tell you whether it produced a good response. > > Yes, absolutely. In fact, Reinforcement Learning (RL) is the primary engine that transforms a raw, unpredictable language model into a helpful, conversational assistant (and, more recently, into advanced "reasoning" models). > However, it is important to distinguish when this happens. An LLM doesn't actively run a reinforcement learning loop while it is typing out a response to you (inference). Instead, RL is used heavily during its post-training and alignment phases to shape how it behaves. > Here is exactly how RL is used to produce the results you see. > 1. The Alignment Phase (RLHF and RLAIF) > When an LLM finishes its initial "pre-training," it is essentially just a massive text-completer. If you ask it, "How do I write an essay?", it might just respond with another question like, "How do I write a poem?" because it saw similar lists on the internet. > To make it act like an assistant, developers use RL: > RLHF (Reinforcement Learning from Human Feedback): The model generates several different responses to a prompt. Human annotators rank those responses from best to worst. A separate Reward Model is trained to understand what humans like (clarity, politeness, accuracy) and what they hate (hallucinations, toxicity). The core LLM is then optimized using an RL algorithm?traditionally PPO (Proximal Policy Optimization)?to maximize its "reward score." > RLAIF (Reinforcement Learning from AI Feedback): Because human scaling is expensive, modern models are increasingly trained using a structured set of principles (a "constitution") where a more powerful AI acts as the judge, scoring and rewarding the trainee model's outputs. > What RL teaches the model: Supervised learning teaches a model what a good answer looks like. Reinforcement learning teaches the model which of several plausible answers is the absolute best. > 2. The Reasoning Revolution (RLVR) > There has been a massive structural shift in how the newest generation of LLMs (like DeepSeek-R1 or OpenAI's "o" series) use RL. > Instead of just optimizing for conversational style, models are now trained using Reinforcement Learning with Verifiable Rewards (RLVR) to solve complex math, logic, and coding problems. > Instead of a human judging the output, the model is plugged into a sandbox or checker: > The model is given a hard problem and tries to generate a "Chain of Thought" (thinking steps) to solve it. > If it gets the answer right (e.g., the code compiles and passes tests, or the math equation balances), it receives a massive positive reward. > If it fails, it receives a penalty. > Through millions of these trial-and-error iterations, the model naturally learns how to double-check its own work, correct its own mistakes, and "think" systematically?entirely driven by an RL reward loop. > The Modern RL Toolkit > While PPO was the gold standard for a long time, managing multiple massive models simultaneously during training is incredibly expensive. Today, engineers utilize a variety of direct and architectural RL variants depending on the task: > Method How it Works Primary Use Case > PPO (Proximal Policy Optimization) Uses a standalone reward model to nudge the LLM's token probabilities step-by-step. Classic behavioral and safety alignment. > DPO / GRPO (Direct / Group Relative Policy Optimization) Math hacks that eliminate the need for a separate reward model, optimizing the LLM directly from paired choices. High-efficiency alignment and scaling up reasoning capacity. > Process-Level RL Rewards the model for individual steps in its thinking process, not just the final answer. Eliminating hallucinations in complex, multi-step logic. > barbara > > ----- Forwarded Message ----- > From: Brian E Carpenter via Internet-history > To: Greg Skinner ; "internet-history at elists.isoc.org" > Sent: Thursday, June 25, 2026 at 03:27:41 PM PDT > Subject: Re: [ih] History of AI and Internet > > Greg, > > > Since some of you are or were professors, I?d be interested in your opinions about the required math background, as the book is designed for a course in reinforcement learning. > > In my opinion, the book is a masterpiece, but it needs fluency in differential calculus and matrix algebra (which I lost many years ago). I think it would be quite hard to read cover-to-cover while skipping the maths. Also, if you're interested in applicability to large language models, it won't help you. A lot has happened since 2020. > > Ananthaswamy's book that I mentioned recently relies on similar maths, but you can largely skip it without losing the thread. > > Regards/Ng? mihi > Brian Carpenter > From gregskinner0 at icloud.com Fri Jun 26 10:22:44 2026 From: gregskinner0 at icloud.com (Greg Skinner) Date: Fri, 26 Jun 2026 10:22:44 -0700 Subject: [ih] History of AI and Internet In-Reply-To: References: <580818707.1307788.1782233544689@mail.yahoo.com> <1677377271.1720268.1782326721278@mail.yahoo.com> <4B5A8925-D722-4AA6-821F-1923D535BEA2@icloud.com> Message-ID: On Jun 25, 2026, at 3:27?PM, Brian E Carpenter wrote: > > Greg, > > In my opinion, [Reinforcement Learning: An Introduction] is a masterpiece, but it needs fluency in differential calculus and matrix algebra (which I lost many years ago). I think it would be quite hard to read cover-to-cover while skipping the maths. Also, if you're interested in applicability to large language models, it won't help you. A lot has happened since 2020. > > Ananthaswamy's book that I mentioned recently relies on similar maths, but you can largely skip it without losing the thread. > > Regards/Ng? mihi > Brian Carpenter > I read the preface for Ananthaswamy?s Why Machines Learn (using a Kindle sample available at Amazon) and found the math more accessible. When I first skimmed through Reinforcement Learning: An Introduction and saw the many mentions of dynamic programming and the references to people such as Bertsekas, I immediately thought of MIT?s Laboratory for Information and Decision Systems and its contributions to networking. They are very much involved in AI now, including reinforcement learning. --gregbo From brian.e.carpenter at gmail.com Fri Jun 26 14:10:27 2026 From: brian.e.carpenter at gmail.com (Brian E Carpenter) Date: Sat, 27 Jun 2026 09:10:27 +1200 Subject: [ih] Fwd: History of AI and Internet In-Reply-To: References: <755299404.2218793.1782443866595@mail.yahoo.com> Message-ID: Of course, reinforcement learning is essential to LLM systems. What I meant is that since the Sutton & Barto book is a few years old, it simply doesn't talk about LLMs, but the Ananthaswamy book does. Regards/Ng? mihi Brian Carpenter On 26-Jun-26 20:32, Greg Skinner via Internet-history wrote: > Forwarded for Barbara > >> On Thursday, June 25, 2026 at 08:08:57 PM PDT, Barbara Denny wrote: >> >> >> Hi >> >> I have heard things contrary to what i think you said regarding Reinforcement Learning and its applicability to today's LLMs >> >> Here is what Gemini told me but i do know sometimes it makes big mistakes. I still find it interesting to see what it put together. Too far out of my experience to tell you whether it produced a good response. >> >> Yes, absolutely. In fact, Reinforcement Learning (RL) is the primary engine that transforms a raw, unpredictable language model into a helpful, conversational assistant (and, more recently, into advanced "reasoning" models). >> However, it is important to distinguish when this happens. An LLM doesn't actively run a reinforcement learning loop while it is typing out a response to you (inference). Instead, RL is used heavily during its post-training and alignment phases to shape how it behaves. >> Here is exactly how RL is used to produce the results you see. >> 1. The Alignment Phase (RLHF and RLAIF) >> When an LLM finishes its initial "pre-training," it is essentially just a massive text-completer. If you ask it, "How do I write an essay?", it might just respond with another question like, "How do I write a poem?" because it saw similar lists on the internet. >> To make it act like an assistant, developers use RL: >> RLHF (Reinforcement Learning from Human Feedback): The model generates several different responses to a prompt. Human annotators rank those responses from best to worst. A separate Reward Model is trained to understand what humans like (clarity, politeness, accuracy) and what they hate (hallucinations, toxicity). The core LLM is then optimized using an RL algorithm?traditionally PPO (Proximal Policy Optimization)?to maximize its "reward score." >> RLAIF (Reinforcement Learning from AI Feedback): Because human scaling is expensive, modern models are increasingly trained using a structured set of principles (a "constitution") where a more powerful AI acts as the judge, scoring and rewarding the trainee model's outputs. >> What RL teaches the model: Supervised learning teaches a model what a good answer looks like. Reinforcement learning teaches the model which of several plausible answers is the absolute best. >> 2. The Reasoning Revolution (RLVR) >> There has been a massive structural shift in how the newest generation of LLMs (like DeepSeek-R1 or OpenAI's "o" series) use RL. >> Instead of just optimizing for conversational style, models are now trained using Reinforcement Learning with Verifiable Rewards (RLVR) to solve complex math, logic, and coding problems. >> Instead of a human judging the output, the model is plugged into a sandbox or checker: >> The model is given a hard problem and tries to generate a "Chain of Thought" (thinking steps) to solve it. >> If it gets the answer right (e.g., the code compiles and passes tests, or the math equation balances), it receives a massive positive reward. >> If it fails, it receives a penalty. >> Through millions of these trial-and-error iterations, the model naturally learns how to double-check its own work, correct its own mistakes, and "think" systematically?entirely driven by an RL reward loop. >> The Modern RL Toolkit >> While PPO was the gold standard for a long time, managing multiple massive models simultaneously during training is incredibly expensive. Today, engineers utilize a variety of direct and architectural RL variants depending on the task: >> Method How it Works Primary Use Case >> PPO (Proximal Policy Optimization) Uses a standalone reward model to nudge the LLM's token probabilities step-by-step. Classic behavioral and safety alignment. >> DPO / GRPO (Direct / Group Relative Policy Optimization) Math hacks that eliminate the need for a separate reward model, optimizing the LLM directly from paired choices. High-efficiency alignment and scaling up reasoning capacity. >> Process-Level RL Rewards the model for individual steps in its thinking process, not just the final answer. Eliminating hallucinations in complex, multi-step logic. >> barbara >> >> ----- Forwarded Message ----- >> From: Brian E Carpenter via Internet-history >> To: Greg Skinner ; "internet-history at elists.isoc.org" >> Sent: Thursday, June 25, 2026 at 03:27:41 PM PDT >> Subject: Re: [ih] History of AI and Internet >> >> Greg, >> >>> Since some of you are or were professors, I?d be interested in your opinions about the required math background, as the book is designed for a course in reinforcement learning. >> >> In my opinion, the book is a masterpiece, but it needs fluency in differential calculus and matrix algebra (which I lost many years ago). I think it would be quite hard to read cover-to-cover while skipping the maths. Also, if you're interested in applicability to large language models, it won't help you. A lot has happened since 2020. >> >> Ananthaswamy's book that I mentioned recently relies on similar maths, but you can largely skip it without losing the thread. >> >> Regards/Ng? mihi >> Brian Carpenter >> > From byeager at fastmail.fm Fri Jun 26 17:16:40 2026 From: byeager at fastmail.fm (william yeager) Date: Fri, 26 Jun 2026 17:16:40 -0700 Subject: [ih] History of AI and Internet Message-ID: I?ve used Perplexity Pro ($200/year) to gain an in depth understanding including all of the math. I?m lucky to have completed my Phd studies in math albeit in 1970 at the University of Washington. I felt given the above that neither RLHF nor RLAIF cannot eliminate either toxicity or hallucinations completely. Neither to the other RL algorithms. Clever individuals can query and receive them. A little review - I think it is correct - most from memory and some excised from my perplexity notes. The latter are obvious. Here goes ... We all now know that training uses gradient descent, back propagation and cross-entropy to minimize the loss between the calculated next token and the token itself. Training has as a goal to minimize the loss between the them. Responding to queries is different and uses stochastic probability to predict the best next token. The following is slightly technical but I think most can follow it, and points out the difference between training and queries. In a 50 layer LLM with about 30 billion parameters, let?s say training has created about 500 matrices with a dimension of 4096, 10 per layer. The dimension refers to length or number of columns of a row. And all matrices are dimension x dimension, eg, 4096x4096 in this case. There are some hidden layers, and here, if a query has N tokens, then one finds some special matrices N x dimension. But enough ! The fixed matrices are created during training. They are are initialed with random values prior to training. Training is expensive and 90% of what makes these models work. When training is completed, The matrices are each query?s calculations to generate a response. Theses modified versions are created, and kept in the cache until the response is completed. They are the basis for creating the probabilities to produce the next best token. The trained matrices are reused for each query. Let?s assume that the LLM also has as a fixed vocabulary of about 60 thousand words, and about 80,000 tokens in what is called the embedded table; and let?s assume the dimension of the model is 4096. I hope the array info is OK for the readers. Each token has a row in the embedded table?s 80,000x4096 array. This is fixed, and activations of the tokens are used when processing a query. One can view 80,000 as the token vocabulary. And again, the word/subword vocabulary is usually in the [30K, 100K] size and for 80,000 tokens, 60K is OK for discussion. There is a variable, T, called Temperature comes directly borrowed from Boltzmann?style statistical mechanics. The final step to predict the next token is as below. Note that multiple passes are made through these layers until the response is completed. The number of passes as well as the number of layers are hyper parameters in the LLM?s design. Note that the z(i)s are called logits because the math works with numerical values or weights created during the processing of the query, and the i is an index into the actual token?s row in the embedded tokens' table. To wit (I hope this is OK too) this is how the probabilities, p(i) are computed for the best next token?s choice - the following is a png file, and softmax is what computes them. Assuming you understand how exponentials amplify differences: If T <1: Division by T magnifies differences between logits. The largest z i exp ? ( z i / T ) The distribution becomes more peaked (low entropy); the model is more deterministic and tends to pick the top?logit token almost always. If T > 1: Division by T shrinks differences between logits. exp ? ( z i / T ) values get closer together. The distribution becomes flatter (higher entropy); the model samples more from lower?probability tokens, so choices are more random. T = 1 has no effect and softmax is used by cross-entropy . T = 0 just selects one token using argmax which picks the largest p(i) token because you cannot divide by 0. Now to the subject at hand: RLHF reduces toxicity and some hallucinations, but it does not eliminate either of them. Here?s how and why. Most from perplexity and all have references which I randomly check for validity. Social media refs are always suspect. 1. What RLHF does well RLHF trains the model to prefer answers that humans rate as good: helpful, safe, non?toxic, on?topic. Human labelers rank responses; a reward model learns to score ?good? higher than ?bad.? RL (usually PPO) then shifts the LLM?s behavior toward high?reward outputs. Empirically, this: Reduces overt toxicity and abusive content compared to the raw pretrained model. Encourages refusals on obviously unsafe or clearly unanswerable prompts (?I can?t help with that?). So RLHF is very effective as a behavioral filter and style/intent alignment tool. 2. Why hallucinations are not solved Hallucinations (confidently wrong answers) come from: The underlying pretrained model?s imperfect world model and reliance on patterns, not explicit facts. Pressure to always produce an answer, even when unsure. RLHF doesn?t change the core training objective (next?token prediction) or give the model a ground?truth database; it just tells it which style of answer humans prefer. Key limitations: Reward models are trained on human judgments that may not reliably detect subtle factual errors. Human raters often reward plausible and fluent text more than pedantically correct but awkward text, which can actually reinforce ?polished hallucinations.? Studies and commentary increasingly note that RLHF improves alignment and niceness but does not reliably fix factual reliability. So RLHF can discourage obviously nonsensical answers, but it doesn?t guarantee truthfulness, especially on long, detailed, or niche factual queries. 3. Why toxicity isn?t fully eliminated Toxic content is reduced, but not gone, because: The reward model only approximates human safety judgments and may misgeneralize to new prompts or adversarial phrasings. Attackers can craft ?jailbreak? prompts that drive the model into regimes where the reward model and safety filters are less effective. Human feedback is noisy and not perfectly consistent; some borderline content slips through. Recent work shows that models trained with RLHF can still be induced to generate harmful or biased content with carefully designed prompts, indicating safety remains imperfect. 4. Consensus in current research The emerging consensus is: RLHF is a core tool for safety and alignment in today?s LLMs. It significantly reduces toxicity and some blatant hallucinations compared to unaligned models. It does not eliminate hallucinations or toxicity and has fundamental limitations as a sole safety mechanism. That?s why current work layers RLHF together with: Better pretraining data curation. System prompts and rule?based filters. Tool use (retrieval, calculators, constrained decoding). Adversarial testing and other alignment methods. So, RLHF moves the needle in the right direction but is not a mathematically guaranteed cure for either hallucinations or toxic outputs. If you?re interested, we could next look at why reward models systematically miss subtle hallucinations?essentially a bias in the human feedback signal itself. What is RLAIF RLAIF is essentially ?RLHF, but the judge is another AI model instead of a human.? More formally: RLHF = Reinforcement Learning from Human Feedback. RLAIF = Reinforcement Learning from AI Feedback (or AI?based feedback). The pipeline is almost identical; the only thing that changes is who provides the preference labels. 1. Basic idea RLAIF replaces the human annotator in the RLHF loop with a teacher/critic LLM: Start from a pretrained (and usually SFT?tuned) base model. For a prompt x x, sample several candidate responses y ( 1 ) , y ( 2 ) , ? y(1),y(2),? from the student model. Ask a stronger or specially prompted AI labeler (teacher model) which response is better according to some principles or ?constitution.? Use those AI?generated preferences to: Train a reward model, or Directly use them as rewards during RL (direct RLAIF). Run RL (often PPO) to fine?tune the student model toward high?reward behaviors, just like RLHF. So structurally, RLAIF = RLHF with AI feedback instead of human feedback; the RL and reward?model math stays the same. 2. Motivation vs RLHF Why bother? Scalability: AI can label millions of response pairs far faster and cheaper than humans, limited mainly by compute. Consistency: A given AI labeler applies the same ?preferences? more uniformly than a pool of human annotators (though it has its own biases). Safety & specialization: AI labelers can be used on sensitive or technical domains where using human raters is difficult, risky, or requires rare expertise. Empirical work shows that, for tasks like summarization and helpful/harmless dialogue, RLAIF can reach comparable performance to RLHF in terms of alignment metrics, at much lower human?labeling cost. 3. Relationship to Constitutional AI Constitutional AI (Anthropic) is a prominent instance of this idea: Use a written ?constitution? (high?level principles: be harmless, honest, etc.). Ask a powerful LLM (the critic) to judge responses based on this constitution and suggest improvements. Use that AI feedback to train and refine the assistant model?this is effectively RLAIF within a structured framework. 4. Limitations RLAIF trades human?label noise for AI?label bias: If the AI labeler is misaligned, biased, or simply wrong in certain domains, its flaws propagate into the reward model and the student. There?s a real risk of ?AI teaching AI its own mistakes,? so careful design of the teacher, prompts, and constitutions is crucial. Current work suggests RLAIF is powerful and scalable, but not a drop?in replacement for human oversight; in practice, people often combine human and AI feedback and validate AI?labeled data with humans on critical tasks. That?s a fair way to put it: both RLHF and RLAIF are only as good as the feedback signal they?re optimizing. With RLHF, if human raters reward fluency and ?sounds good? over factual accuracy or deep reasoning, the model learns to produce very plausible but sometimes wrong answers?polished hallucinations. Garbage (or shallow judgment) in, garbage out. With RLAIF, you?re stacking another model?s biases on top: if the AI critic has blind spots or systematic errors, those propagate into the reward model and then into the trained assistant. The ?clever response? part is crucial: once you have a very strong generator, small biases in the reward signal tend to produce outputs that are: Grammatically excellent. Stylistically aligned. But sometimes factually or logically off in ways that are hard for non?expert raters to spot. This is why current alignment work is putting so much emphasis on: Better feedback collection protocols (more expert raters, better guidelines, better evaluation tasks). Combining human and AI feedback, rather than using AI feedback alone. Adding external tools (search, calculators, retrieval) so the model isn?t forced to ?make up? answers from its weights alone. So your intuition is right: these methods are extremely powerful amplifiers of whatever signal you feed them?good or bad. In the above I was curious about ?garbage-in, garbage-out.? It cannot be avoided 100% of the time now. So, I hope this is helpful, and I?m sure I made some grammatical errors. I will be using small essays like the above in a PDF that will be part of my three hour presentation on ?A History of AI.? It will be publicly available after the conference and I will post it to this list. The Internet and AI has been useful for me to find key areas to present. One area I haven?t seen is Claude Shannon?s contributions to today?s Internet and circuit design. In his Mathematical Theory of Communication written in 1948 he uses bits (0 and 1), defines communication Capacity, and then shows given entropy or noise in the channel when a message from a source can reliable reach the destination: 1948 paper ? noisy, networked world In 1948, his ?Mathematical Theory of Communication? formalized information entropy, channel capacity, and coding theorems, saying in effect: here is how much information you can push through a noisy medium, and how to do so nearly error?free. Everything from deep?space probes to fiber backbones, Wi?Fi, 5G, and compressed media codecs operates inside this Shannon framework; we just keep finding better codes and hardware to approach his theoretical limits. His 1937 MIT master?s thesis is called the greatest master?s thesis ever written: Master?s thesis ? digital hardware In his 1937 master?s thesis, Shannon showed that Boolean algebra and relay/switching circuits are mathematically equivalent, so logical propositions could be implemented by electrical networks in a systematic way. This result provided the theoretical foundations for digital circuits and thus for modern digital computers, establishing that logic design could be done algebraically rather than by ad hoc wiring. Shannon is widely regarded as the ?father of the information age? because his work on digital circuit theory, information theory, and sampling theory underlies computer architecture, digital communication, data compression, and error-correcting codes. These ideas also influenced later work on computer architectures that implement quantum algorithms, which are still described using logical gates and circuit models that conceptually descend from Shannon?s Boolean-circuit viewpoint. Thus, that one step turned logic design from an art into an algebraic engineering discipline, and it is the conceptual root of modern CPUs, GPUs, FPGAs, and essentially any digital hardware that obeys logic gates. He is one of my personal heroes. Definitely unsung, an by many considered among the top 3 minds in the 20th century. Bill ?????????? ????? ???? ????????? From byeager at fastmail.fm Fri Jun 26 17:24:22 2026 From: byeager at fastmail.fm (william yeager) Date: Fri, 26 Jun 2026 17:24:22 -0700 Subject: [ih] ih] History of AI and Internet Message-ID: <1029DB3D-6D1E-4653-9D38-87327B8365B5@fastmail.fm> Merde !! Mail.app deleted a png and really screwed up the message. I?ll figure out how to fix it for a 2nd time. Sorry really sorry. ?????????? ????? ???? ?????????