They are new proofs and for sure useful but as far as I’ve understood mainly interpolative. E.g. an LLM can “create” a poem about a purple chicken as it has datapoints for “purple” and “chicken”, so it can create something plausible inbetween.
Similarly, in my mind it can interpolate proofs by interpolating between data points for technique A and technique B. This is novel and brute-forcing proofs this way is useful. It is analogus to how sometimes it can generate programs that pass unit tests, I think.
However, creating fundamentally new concepts outside of the interpolated datapoints is not something I am convinced of. Maybe it can extrapolate some things, if correct add it as a data point, continue. Essentially a search, and it would be amazing if this works and maybe we can get some recursive improvement this way. But the “ideas” it will use to conduct this search are a function of the input data points as well, and thus in my view fundamentally limited in novelty. I am not discounting the usefulness, but I am not convinced you can just keep doing this indefinitely scaling intelligence exponentially.
Of course nobody can know yet really and I am just speculating just like you. But I also think the “experts” Sam and Dario also don’t know, and given their incentives I am not really convinced by them.
They're finding new examples in well-known categories of stuff. Also proving is a question of search. They're finding some stuff we missed, but not because they're smart, just because they have lots of data and compute. Think of them as something closer to brute force.
>Think of them as something closer to brute force.
Well brute force is something I had waited for all my life so it could make a personal computing experience higher performance and more fulfilling.
By the time 1980 rolled around.
In some ways it's like a fundamental assumption that's always been there and nobody questions or really "thinks" about. It hurts the brain, sooner or later you want a computer to be able to do some thinking for you ;)
It's one of the easiest tech concepts to understand, especially for those who don't have very deep abilities.
Hardware is hardware, stronger is stronger, and quantity has its own qualities.
People in certain positions can go farther with nothing more than this firm a grasp of computer science or intelligence and it shows.
The only thing to be gained by continuing to grow the scale of "datacenters" from this point is the brutishness.
I expect it to pay off too, for those who can afford it, just not for everybody.
What astonishes me about singularity skepticism: every argument was, from when it was released, "but it can't do X", "it's impossible/too expensive". The benefits AI -> AGI are staggering, the competition is intense (both intra-industry and internationally), and the gains self-reinforcing.
I have fears of dystopian worst-case scenarios, and dread the rapid pace of change and what it will do to real lives lived in the world, but it's only clanker cope/wishcasting that believes that (AI+human ingenuity) won't produce real AI ingenuity.
Re-read your comment in four years, I'll take the bet you'll find it very naive.
Similarly, in my mind it can interpolate proofs by interpolating between data points for technique A and technique B. This is novel and brute-forcing proofs this way is useful. It is analogus to how sometimes it can generate programs that pass unit tests, I think.
However, creating fundamentally new concepts outside of the interpolated datapoints is not something I am convinced of. Maybe it can extrapolate some things, if correct add it as a data point, continue. Essentially a search, and it would be amazing if this works and maybe we can get some recursive improvement this way. But the “ideas” it will use to conduct this search are a function of the input data points as well, and thus in my view fundamentally limited in novelty. I am not discounting the usefulness, but I am not convinced you can just keep doing this indefinitely scaling intelligence exponentially.
Of course nobody can know yet really and I am just speculating just like you. But I also think the “experts” Sam and Dario also don’t know, and given their incentives I am not really convinced by them.