No article about vintage AI is complete without mentioning the Dartmouth summer workshop of 1956, which is widely regarded as the event that kickstarted the entire field of AI research. Participants included people like John McCarthy, Marvin Minsky, Ray Solomonoff, Claude Shannon, and John Nash, among others. In retrospect, they might have slightly underestimated the nontriviality of the problems involved:
> An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.
I mean, it took a bit more than eight weeks, and a few doublings of hardware capacity, but these days we have obviously made tremendous progress in making machines "use language, form abstractions and concepts, [and] solve kinds of problems now reserved for humans". The self-improvement part is – perhaps luckily – trickier, although what else are generative adversarial networks if not that? AlphaGo rapidly self-improved itself to superhuman levels without any human help after a certain point!
Stuff like checkers was a "a kind of problem reserved for humans" but was solved decades ago; you can see why the situation is similar to the context of the above quote. Time and time we have been through claims of "X solved, we are just a summer away from thinking machines" ... yet after the smoke cleared we only had a better solver for X, and nothing else. We may have a ridiculously better solver for chess, but that's about it -- it's as far from "abstracting and forming concepts" and "self-improving" as everything from the 60s was (in retrospect).
For me the biggest difference is that the current days we are happy with more black boxy methods where it's unclear if they can lead to actual progress or applicability.
Or it’s that as computers learn to solve problems previously reserved for humans, people tend to shift the goalposts of what counts as real intelligence accordingly, similar to the so-called "God in the gaps" fallacy. (Incidentally, something very similar has been going on as our understanding of animal intelligence has improved.)
That said, I agree that however impressive narrowly intelligent tasks computers have learned to excel at, most of the research involved has not helped us make much (if any) headway towards figuring out how to create a fully general machine intelligence despite the early optimism.
But that said, for the past several decades the mainstream AI research community has mostly not even attempted to pursue AGI research, being entirely content with working on different types of narrow AI and indeed trying to dispel any illusions that their work might have anything to do with "science fiction AIs". This, of course, exactly because of those broken early dreams and the AI winter that followed.
So, at least within the field itself, I don’t really agree that there has been any tendency to assume that narrow AI breakthroughs imply that AGI is just around the corner. The current hype around GANs and LLMs is an exception, not the rule.
I have been told since before I ever touched a computer (70s) that programmers would soon not be needed because you can just tell your computer what to build using AI. And then that AI would destroy the world in the cold war of course.
> The word robot comes from the word robota. The word robota means literally "corvée", "serf labor", and figuratively "drudgery" or "hard work" in Czech.
> ... this whole debate is reminiscent of Karel Čapek's War with the Newts [1936]. In particular the public discourse from a time before the newts took over. "It would certainly be an overstatement to say that nobody at that time ever spoke or wrote about anything but the talking newts. People also talked and wrote about other things such as the next war, the economic crisis, football, vitamins and fashion; but there was a lot written about the newts, and much of it was very ill-informed. This is why the outstanding scientist, Professor Vladimir Uher (University of Brno), wrote an article for the newspaper in which he pointed out that the putative ability of Andrias Scheuchzer to speak, which was really no more than the ability to repeat spoken words like a parrot, ..." Note the irony of the professor's attempt to improve an ill-informed debate by contributing his own piece of misinformation, equating newt speech to mere parrot-like mimicry.
The idea of automated mechanization goes much, much further back, at least as far as Proto-Indo-European language (ca. 4500 - 4000 BC), predating even the Greek story of Talos. See:
People seem to forget that the moral of "The Boy Who Cried Wolf" is that the wolf does finally come. It's not "don't panic prematurely because people will stop listening to you"; it's "don't disbelieve real hazards even after repeated false alarms".
If I can jump in here: it's something like "be very careful when making extreme claims, because they probably won't be believed. If your claims turn out to be false, then they really won't be believed next time."
So it's about trust and reputation. If the village elder, whom everyone trusts, cried "wolf!" the villagers would be inclined to believe him.
If you keep lying or making false alarms, people won't believe you when you're actually in trouble. So, it's better to tell the truth and be trustworthy, or you'll end up losing the trust of others when you really need it.
I think the moral is about the boy, not the villagers and what they believe - it's that reputation is an important thing, because once you damage yours, you won't be able to use it even when you really need it.
Don't have a small child be responsible for protecting your sheep from wolves? Esepcially if you have already determined not to come if he shouts for help.
"There were 1,000,000 different doom-prophets in this civilization. 1 of them knew what she was talking about. The people mocked the doom-prophets. Doom-mocking was a safe career, 99.9999% of the time they were correct. Ruin came for them all the same."
People making doomsday predictions are always going to claim that _their_ prediction is the “real hazard” unlike all those other guys. Of course we should dismiss everyone else’s doomsday claims, those are just silly; but it would be a horrible folly to dismiss mine.
For centuries man had attempted to fly, and for centuries his friends told him it would never happen or that it would only happen millions of years into the future. They called him a crank and a kook and a charlatan. Some were still calling him that on 17 December 1903.
"Climate predictions show our hopes and fears aren't new, even if more CO2 is being released into the atmosphere than ever before"
What I will say is that people probably weren't wrong to be concerned previously. I wrote a comment yesterday about how ChatGPT is probably more conscious than a rock. In a similar way I'm fine saying that a computer "thinks" far more than most collections of matter we find in the universe, and I'll add that their ability to "think" seems to be increasing every year.
When computers were first created there was no trend in progress to point to. We knew we could create computers that did some basic "thinking", but when would this become competitive with human ability?
Only now in hindsight can we say the jump from digital computers to AGI is a significant one. But this isn't always the case – a good example being the time between the first flight to landing on the moon. It was perhaps reasonable to think the progress from digital computers to AGI could have been equally as rapid.
33 comments
[ 3.0 ms ] story [ 81.3 ms ] thread> An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.
For me the biggest difference is that the current days we are happy with more black boxy methods where it's unclear if they can lead to actual progress or applicability.
Everything else is a nice side quest. https://arxiv.org/abs/2212.11281
That said, I agree that however impressive narrowly intelligent tasks computers have learned to excel at, most of the research involved has not helped us make much (if any) headway towards figuring out how to create a fully general machine intelligence despite the early optimism.
But that said, for the past several decades the mainstream AI research community has mostly not even attempted to pursue AGI research, being entirely content with working on different types of narrow AI and indeed trying to dispel any illusions that their work might have anything to do with "science fiction AIs". This, of course, exactly because of those broken early dreams and the AI winter that followed.
So, at least within the field itself, I don’t really agree that there has been any tendency to assume that narrow AI breakthroughs imply that AGI is just around the corner. The current hype around GANs and LLMs is an exception, not the rule.
> The word robot comes from the word robota. The word robota means literally "corvée", "serf labor", and figuratively "drudgery" or "hard work" in Czech.
https://en.wikipedia.org/wiki/Karel_Capek#Etymology_of_robot
> R.U.R. is a 1920 science-fiction play by the Czech writer Karel Čapek. "R.U.R." stands for Rossumovi Univerzální Roboti
https://en.wikipedia.org/wiki/R.U.R._(Rossum%27s_Universal_R...
So a century, at least.
This time it's not sci-fi.
> ... this whole debate is reminiscent of Karel Čapek's War with the Newts [1936]. In particular the public discourse from a time before the newts took over. "It would certainly be an overstatement to say that nobody at that time ever spoke or wrote about anything but the talking newts. People also talked and wrote about other things such as the next war, the economic crisis, football, vitamins and fashion; but there was a lot written about the newts, and much of it was very ill-informed. This is why the outstanding scientist, Professor Vladimir Uher (University of Brno), wrote an article for the newspaper in which he pointed out that the putative ability of Andrias Scheuchzer to speak, which was really no more than the ability to repeat spoken words like a parrot, ..." Note the irony of the professor's attempt to improve an ill-informed debate by contributing his own piece of misinformation, equating newt speech to mere parrot-like mimicry.
> http://gutenberg.net.au/ebooks06/0601981h.html
The idea of automated mechanization goes much, much further back, at least as far as Proto-Indo-European language (ca. 4500 - 4000 BC), predating even the Greek story of Talos. See:
* https://en.wikipedia.org/wiki/H1n%CC%A5g%CA%B7nis
* https://en.wikipedia.org/wiki/Talos
I researched this for a scene in my novel. (Alpha readers wanted, see profile for details.)
If you trace the etymology of Hephaestus, the creator of Talos, you eventually end up at h1n̥gʷnis, if I recall correctly.
So it's about trust and reputation. If the village elder, whom everyone trusts, cried "wolf!" the villagers would be inclined to believe him.
The moral to the listener is to not be like the boy. The moral is not, to be unlike the villagers.
TL;DR: Don't lie, or trust will be lost.
What I will say is that people probably weren't wrong to be concerned previously. I wrote a comment yesterday about how ChatGPT is probably more conscious than a rock. In a similar way I'm fine saying that a computer "thinks" far more than most collections of matter we find in the universe, and I'll add that their ability to "think" seems to be increasing every year.
When computers were first created there was no trend in progress to point to. We knew we could create computers that did some basic "thinking", but when would this become competitive with human ability?
Only now in hindsight can we say the jump from digital computers to AGI is a significant one. But this isn't always the case – a good example being the time between the first flight to landing on the moon. It was perhaps reasonable to think the progress from digital computers to AGI could have been equally as rapid.