omw_68
6hr ago
It's trivially easy to show that you can't.
for x := 2 to infinity for y := 1 to 2x if y is 2x, then halt if isprime(y) and isprime(2x-y) then break
If you can tell whether this program will halt or not, Goldbach would like a word with you.
Very good point, thank you for posting this.
omw_68
7hr ago
I'm not convinced the human element definition is wrong, even if I wouldn't fully endorse it today either.
Ok, well suppose someone wants to go to a bar and have their drink served by a real human bartender. By definition, that's not a task which can be replaced by AI.
Demonstrably, together we got to where we are now. It'd be interesting to see how well AI could start from scratch.
I would guess that if you gave it hundreds of millions of years, it could do pretty well. I should probably note that this issue is kind of academic. I mean, if we stipulate for the sake of argument that development of AI is dependent on human input, it doesn't really say much about what AI could or couldn't do in the future -- standing on the shoulders of giants.
omw_68
8hr ago
All the subjective things? Enjoy a beautiful sunrise?
Well I think it depends on how you define the word "enjoy." In any event, in evaluating AI's capabilities, I think it's best to focus on inputs and outputs. There's really no way to objectively verify an AI's subjective feelings (actually there's no way to objectively verify other people's subjective feelings.)
Raise a child? Be a friend? Meaningful, not just aesthetic art?
I'm pretty confident that AI will absolutely be able to do all the tasks involved in these things. Unless of course you want to define these activities so as to require a human element.
Beyond that, all existent AI I'm aware of was trained on existing human knowledge. Without a human-derived training corpus, what is it capable of?
Without a human-derived training corpus, what is any of us capable of?
omw_68
8hr ago
Has any AI skeptic ever said "I was fundamentally wrong"?
That's an interesting question. As usual, people tend to be far more interested in treating their arguments like soldiers than developing an accurate map of reality.
Big Yud had an interesting point about these kinds of debates. He predicted that at the beginning, there is relatively little debate over what's around the corner in AI. Because what's happening is pretty clearly not human level intelligence. Next, those debates get more intense because AI is developing and the situation is more ambiguous. Finally, those debates reach a crescendo and start dropping off again as it becomes more and more clear that we are just about there.
My sense is that we are more or less over the hill. I feel that AI skeptics like Cal Newport are putting less and less energy into explaining away the achievements of AI and focusing more on general criticism of the technology.
omw_68
8hr ago
I think it's reasonable to point to those as bounds on "intelligence" (which is itself poorly defined): if your super intelligence model is effectively "omniscient", that part is provably not possible.
Yeah, some people seem to think that a smart human can, in general, analyze a computer program and reliably determine if it will halt or not. I'm almost certain that this is not the case.
I think people are used to looking at computer programs which are well organized and therefore it's possible to construct an accurate model of the program and use that model as a shortcut in determining whether the program will halt. So for example, you might study a program; realize that it outputs all prime numbers from 2 on up, and use your knowledge that there is an infinite number of primes to conclude that the program will not halt.
I won't be surprised if it's "less than Skynet" even if it still changes nearly everything.
Yeah, I think these arguments are pure cope. Even if there is something humans can do which AI fundamentally cannot do (doubtful) the technology will almost certainly be wildly transformative.
omw_68
12hr ago
We do not know what resources are required to scale intelligence further up. This is a known unknown.
As an illustration, look at airplane speed records over time. From 1900 to 1945, they rose linearly to the speed of sound. Then between 1955 and 1965, they exploded to 925 m/s. A decade later, the SR-71 offered a modest improvement to almost a kilometer a second. That was 50 years ago, and no new records have appeared since. (Yes, I am skipping rocket planes here, their fuel economy makes supersonic jets look efficient, which is why they remain a curiosity. As are scramjets.)
Well there's at least one thing we do know, which is that the fastest planes are far faster than birds. In other words, it's reasonable to assume that even in the absence of extraordinary progress, AI will end up way smarter than the smartest people.
It seems entirely plausible that the only gain from going from Claude 5.5 to Claude 20.2 will be in the confidence with which the latter will reject the practicality of this.
I agree, there's really no way of knowing what technologies can be discovered and implemented with additional intelligence. Possibly all the low hanging fruit is gone; the middle hanging fruit is mostly gone; and there are a few more developments left before we plateau.
But even so, it's worth noting that AI which is smarter than people in the same way that a car is faster than a Cheetah would have a massive massive impact on the world.
omw_68
12hr ago
But it takes a higher IQ person to come up with superficially insightful but wrong/irrelevant points that leads to true AI denialism (as opposed to AI opposition). Stochastic parrots, Chinese rooms, explanations of why a trend has to stop. They've heard of P =! NP; surely that prevents AI!
Yeah, I've met people who insist that things like the Halting Problem and Godel's Incompleteness Theorem preclude the possibility of AI.
omw_68
13hr ago
So, I'm coming back to this late, and I'm a bit confused by all the multiple threads going on. I think what we're arguing about is basically this: is it racial discrimination if you set up some objective standard, which is going to inherently impact some races more than others? I'm not a lawyer, but as I understand it, it depends on the situation. It's OK for a business to charge high prices knowing that there's a racial wealth gap. It's not ok to charge a fee to vote, even $1, because voting is a human right. It's... questionable for schools to charge high tuition, and they're highly encouraged to give out scholarships to black kids to balance things out. Similiarly, while the big tech companies don't have to hire black people, they try really hard to balance things out so they don't look bad. I'm not a lawyer, but my understanding is that this is a tricky subject, with lots of confusing court decisions all over the case. Is that fair?
I think you are muddling a lot of issues here, but I'm pretty confident that generally speaking if there's an objective standard, and that objective standard is bona fide and legitimate, there's no discrimination problem. So for example, if a private club genuinely needs a lot of money to operate, it's high fees won't be seen as discriminatory even if they have a disparate impact. Similarly, if a track team recruits only sprinters who can run 100 meters in less than 10 seconds, I highly doubt that it's going to be a problem, even if the requirement results in a racial disparity. To be clear, I am going strictly by common sense here.
omw_68
14hr ago
Some concepts are just too high IQ for most people.
I agree, but I think there's another issue here, which is that most people are very conventional in their thinking. They are told that they should be concerned about the Latest Thing and they follow suit with everyone else in their circles.
In any event, I agree that full on GAI (followed shortly by ASI) is just around the corner if it's not here already.
Another factor to keep in mind is that almost certainly there are multiple AI Manhattan Projects going on all over the world right now. Because defense planners are surely aware of the military implications of this technology. I would guess 2-3 in the US; 1-2 in Europe; 1-2 in China; and 1 in Israel.
So that's 5-8 AI projects which are well-funded, well-staffed, etc. and it seems to me that there is a good chance that at least one of these projects is ahead of the public performance of OpenAI, Anthropic, etc. In fact, I wouldn't be surprised if GAI already exists in some lab in Maryland.
omw_68
16hr ago
Uh, people do pretend to be retarded to win sports competitions. Here you go.
I didn't have time to watch the whole video, but it seems to relate to the Paralympics, not the Special Olympics. Am I nitpicking here? I don't think so, because the video says right at the start (about 15 seconds in) that the Paralympics got a lot of attention: "The entire country [Spain] is, needless to say, thrilled."
And here's what the AI said:
Fernando Martín Vicente, president of the Spanish Sports Federation for People with Intellectual Disability (FEDDI), orchestrated the scheme to guarantee gold medals. Winning medals directly unlocked larger state grants, government subsidies, and private corporate sponsorships for the federation.
According to reports in The Smart Set and The Guardian, Vicente held immense power as both a Spanish sports executive and a member of the International Paralympic Committee's ruling board. Delivering consistent Paralympic podium finishes boosted his political standing and kept substantial public funds flowing into the organizations he controlled.
If wins in the Special Olympics produced this kind of attention, status, and money, then yeah, I'm pretty sure that people would pretend to be mentally retarded in order to compete. As far as I know, the Special Olympics doesn't get this level of attention, but feel free to correct me.
omw_68
16hr ago
It's somewhat true already: they get trophies and podium time and even press coverage in excess of much faster athletes who place outside of the top three (prize money may go a bit further, but not to say the 100th finisher). "Local athlete won the [not young] age group in [big marathon]" is a pretty regular headline in local news in my experience. The oldest finisher sometimes gets coverage too.
I'd like to see examples of 60+ runners demanding social status, respect, and/or money commensurate with that of top runners overall. And those demands being taken seriously.
omw_68
16hr ago
There's men's college football and basketball. Why should these get the respect and money they do when at least 60% of the athletes involved couldn't crack a CFL or NBA Europe roster?
Because there is a lot of organic grass roots interest. People will genuinely invest significant time and money to watch these sports. To put it another way, these sports EARNED respect and money.
One can contrast that with (most) women's sports, where respect and money comes from pressuring lawmakers and other elites into putting a big thumb on the scale.
Why do I occasionally hear equally ridiculous things like the idea that the college championship team could beat the Browns or whatever the consensus-worst NFL team is?
What would be (almost) equally as absurd is if a commentator seriously proposed that a particular college team was the best football team ever, including NFL teams. If a commentator proposed that Serena Williams could beat the lowest ranked male players on the Tour, that commentator would probably be wrong but it wouldn't be absurd, not absurd like suggesting that Serena Williams was the best tennis player ever including men.
Long before the current AI race, the NSA was the biggest employer of mathematicians and the biggest consumer of computer equipment in the United States. Probably they have thousands of computer programmers. From what I understand, NSA employees are allowed only to tell their spouses who their employer is. According to AI, the total NSA workforce is about 30,000 people which is far larger than Anthropic.
To be sure, not everyone at the NSA is working on AI, but at the same time I'm pretty confident that the NSA has the manpower to have a Manhattan AI Project going on right now.
Even Unit 8200 in Israel has a much larger workforce than Anthropic.
Presumably there are equivalent organizations in China and Europe.
More options
Context Copy link