Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

languagelog.ldc.upenn.edu

Post reply on HN
Print view View original

11–20 of 55 posts

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#11

Post by ziofill »

dwoldrich wrote: post #5
NitpickLawyer wrote: post #3

> Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. That's a pretty early definition of what we now call ASI (Artificial Super Intelligence). In the next paragraph the author goes to describe what we today call the "singularity" (ASI designing better ASI). But that term seems to be associated to some very weird communities, so the c…

> Well, that didn't happen. The military had stealth aircraft in the 70's. I'll bet they had LLM or better in the 80's and the tech we have now is the consumer-grade version they seeded into industry in the 2020's.

I’ll take that bet

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#12

Post by NitpickLawyer »

Nevermark wrote: post #7
NitpickLawyer wrote: post #3

> Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. That's a pretty early definition of what we now call ASI (Artificial Super Intelligence). In the next paragraph the author goes to describe what we today call the "singularity" (ASI designing better ASI). But that term seems to be associated to some very weird communities, so the c…

>> It is more probable than not that, within the twentieth century, an ultraintelligent machine will be built and that it will be the last invention that man need make, since it will lead to an “intelligence explosion.” > Well, that didn't happen. Predictions are not false because their claims are not in the past yet. That bar would make all predictions wrong. The faster things happen, the higher the speed-of-progres…

> Not proof, but a high bar for evidence that we may be in that explosion now.

I agree, that's why I said "we're seeing things we have working towards this". I think that the "jagged intelligence" that is often used to describe our curent models is confusing a lot of people. On the one hand you have models failing basic "trick" questions that a 5yo would get, but on the other hand you also have models that write better/faster kernels that can run faster/better models, and are currently used to improve the next generation (via dataset prep, filtering, RL environment creation/curation and so on).

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#13

Post by nialse »

NitpickLawyer wrote: post #3

> Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. That's a pretty early definition of what we now call ASI (Artificial Super Intelligence). In the next paragraph the author goes to describe what we today call the "singularity" (ASI designing better ASI). But that term seems to be associated to some very weird communities, so the c…

> > The required high degree of connectivity might be attained with the help of microminiature radio transmitters and receivers.

> Hahaha, this is straight out of 60s-70s sci-fi, where their best futuristic interfaces were smaller CRT screens / flashy keys, etc.

Actually that is not far off from how high bandwidth interconnects work, but through other media.

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#14

Post by bsenftner »

dwoldrich wrote: post #5
NitpickLawyer wrote: post #3

> Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. That's a pretty early definition of what we now call ASI (Artificial Super Intelligence). In the next paragraph the author goes to describe what we today call the "singularity" (ASI designing better ASI). But that term seems to be associated to some very weird communities, so the c…

> Well, that didn't happen. The military had stealth aircraft in the 70's. I'll bet they had LLM or better in the 80's and the tech we have now is the consumer-grade version they seeded into industry in the 2020's.

I had research lab exposure to what the US military had in the 80's, the groups I saw were hyper focused on 3D visualizations in support of in-the-field strategy. Think wireframes superimposed over a live camera feed, with those wireframes being the horizon line and what are the items beyond their current range of view. A building wall is simply a wireframe, and the floors and every type of sensor they could get to feed into visualizations to see inside and past buildings, hills, any navigable area.

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#15

Post by NitpickLawyer »

dwoldrich wrote: post #5
NitpickLawyer wrote: post #3

> Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. That's a pretty early definition of what we now call ASI (Artificial Super Intelligence). In the next paragraph the author goes to describe what we today call the "singularity" (ASI designing better ASI). But that term seems to be associated to some very weird communities, so the c…

> Well, that didn't happen. The military had stealth aircraft in the 70's. I'll bet they had LLM or better in the 80's and the tech we have now is the consumer-grade version they seeded into industry in the 2020's.

There's 0 chance of that. You can maybe say that for things like material science, nuclear stuff, weird physics, and so on (basically anything relating to making big booms, delivering big booms, or ensuring others can't make big booms). But having LLMs as we understand them now in the 80s would be impossible. They simply didn't have the compute necessary. The entire world combined didn't have it.

For reference, a single 4090 GPU has more FP8 flops than the top supercomputer in 2007. A 4x 4090 computer (something you can buy today for ~10k) would be better than to top supercomputer in 2010 [1]. There's a reason deep learning only started to really work in the past few decades. We had the "theory" for a while, but no compute to actually put it in practice. And the current models are being trained on 10s to 100s of thousand of enterprise GPUs, that make the 4090 look like a toy.

[1] - https://timdettmers.com/2023/01/30/which-gpu-for-deep-learni...

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#16

Post by dwoldrich »

not-a-llm wrote: post #6
dwoldrich wrote: post #5

Earlier quoted context omitted.

> Well, that didn't happen. The military had stealth aircraft in the 70's. I'll bet they had LLM or better in the 80's and the tech we have now is the consumer-grade version they seeded into industry in the 2020's.

the same military is using computers with Windows and Xbox controllers and got upset when Anthropic didn't want to let them use their LLM

I'm not saying the movie War Games was a limited hangout for a super-intelligent secret military computer system/network that already existed. I'm just saying it could have been.

I see those trillion dollar data centers we're building in the US are classed as military installations or at least appear to be a public-private partnerships where the data centers are being built on air force bases.

Perhaps the consumer compute tech has advanced to the point that it's good enough to host the next phase of intelligence for the military. Maybe all the water AI datacenters need is for energy, not cooling.

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#17

Post by convolvatron »

dwoldrich wrote: post #5
NitpickLawyer wrote: post #3

> Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. That's a pretty early definition of what we now call ASI (Artificial Super Intelligence). In the next paragraph the author goes to describe what we today call the "singularity" (ASI designing better ASI). But that term seems to be associated to some very weird communities, so the c…

> Well, that didn't happen. The military had stealth aircraft in the 70's. I'll bet they had LLM or better in the 80's and the tech we have now is the consumer-grade version they seeded into industry in the 2020's.

that's really unlikely. in the 80s the US military had several AI labs and was working with academia...I worked with some. the focus was nearly exlucsively on logic programming and 'expert systems', a path which failed miserably leading to an 'AI winter'.

somewhere around 1990, working at a Dod lab, I picked up on the 'neural network' craze and did a little work. it seemed interesting, but not that much different than using adaptive filters for signal processing. that seemed to be the general consensus, that there wasn't a lot of of interesting behavior or depth there. the number of weights and sizes of super computers at the time were a good 3-4 decimal orders of magnitude lower than what's being thrown today.

so unless the military ran a giant expensive program just to hide the fact that they actually had machines 10000x faster than the $50M (in 80s money) supercomputers of the day, and had trained them on a corpus they whipped up out of nowhere (the amount of information on the internet at the time could probably have fit on a few of todays phones), it probably didn't happen.

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#18

Post by not-a-llm »

applfanboysbgon wrote: post #9
not-a-llm wrote: post #6

Earlier quoted context omitted.

the same military is using computers with Windows and Xbox controllers and got upset when Anthropic didn't want to let them use their LLM

While I think GP comment was a bit looney, the first sentence is a fairly weak counterargument. Xbox controllers especially are used specifically for familiarity purposes. The US military can and does produce significantly more advanced control planes, but they aren't necessary for all purposes and if something is simple enough it's easier to not have to train people on something new. Likewise creating an own general…

sure, I was just talking about this Hollywood myth that the military is far ahead of everyone else, including in computing matters when we all know the military procurement birrocracy

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#19

Post by dwoldrich »

NitpickLawyer wrote: post #15
dwoldrich wrote: post #5

Earlier quoted context omitted.

> Well, that didn't happen. The military had stealth aircraft in the 70's. I'll bet they had LLM or better in the 80's and the tech we have now is the consumer-grade version they seeded into industry in the 2020's.

There's 0 chance of that. You can maybe say that for things like material science, nuclear stuff, weird physics, and so on (basically anything relating to making big booms, delivering big booms, or ensuring others can't make big booms). But having LLMs as we understand them now in the 80s would be impossible. They simply didn't have the compute necessary. The entire world combined didn't have it. For reference, a sin…

Why are you so certain you know what tech existed in the 80's?

NASA used Commodore 64's as part of the constellation of computers needed to launch the space shuttle. That whole program was pretty hokey. That silliness doesn't preclude some radically advanced compute tech from existing elsewhere at the same time. Of course something like that would be ultra secret and you and I wouldn't be told about it.

You have to be told something before you can believe it, right? You also get to choose whether you believe that something you're told is 100% god's honest truth or a load of bullshit as well. It's kindof sad to see people try to gatekeep thought here. Lot of rigid thinkers.

I recall Intel in the 90's trying to build clockless pentiums. They couldn't figure it out or maybe it had integration problems with clockful peripherals and it never came to fruition. But I wondered if that was some furtive attempt to seed some tech into industry. I just try to keep an open mind to these sorts of things.

Re: Speculations Concerning the First Ultraintelligent Machine (1965) [pdf]

#20

Post by AnimalMuppet »

Nevermark wrote: post #7
NitpickLawyer wrote: post #3

> Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. That's a pretty early definition of what we now call ASI (Artificial Super Intelligence). In the next paragraph the author goes to describe what we today call the "singularity" (ASI designing better ASI). But that term seems to be associated to some very weird communities, so the c…

>> It is more probable than not that, within the twentieth century, an ultraintelligent machine will be built and that it will be the last invention that man need make, since it will lead to an “intelligence explosion.” > Well, that didn't happen. Predictions are not false because their claims are not in the past yet. That bar would make all predictions wrong. The faster things happen, the higher the speed-of-progres…

>>> It is more probable than not that, within the twentieth century, an ultraintelligent machine will be built and that it will be the last invention that man need make, since it will lead to an “intelligence explosion.”

>> Well, that didn't happen.

> Predictions are not false because their claims are not in the past yet.

"Within the twentieth century" places it definitely in the past, and definitely false.

Post reply on HN