All novel insights expressible as text can be assembled from snippets of existing text. The challenge is "only" that you need to find the right snippets. Hence being a stochastic parrot does not represent an upper bound on capabilities, but a suggestion that if you want your parrot to give you a specific answer, you should arrange for a context where it's statistically likely to be correct.

>It’s going to narrow in on a statistically better answer but it’s not going to create novel insight beyond the arguments made from compressed snapshots of existing knowledge.

ignoring the fundamentally dishonest representation of LLMs in this subthread..

if humanity was given a tool that could restructure already existing and known knowledge and facts into novel results with a statistically better answer automatically it wouldn't be as a minor a win as you make it sound; it would be tremendous.

That's not what "attention" means.

> it’s not going to create novel insight beyond the arguments made from compressed snapshots of existing knowledge

So all of the maths breakthroughs were not novel?

> It’s great at bread and butter stuff, but not at specialised ideation.

This is fundamentally different from your earlier, stronger claim that - "The very best LLM can answer is the answer you can expect from an online forum like Reddit or a book/paper published more than a year ago."

I use a less popular java framework quite a lot, Micronaut (I want to move away from it, but that's a different discussion).

When trying to solve problems with its behaviour, I can watch as Claude digests the framework source to come up with suggestions, identify bugs that are causing behaviour regressions, and generally unlock stuff that I've previously found very frustrating. Frustrating because the documentation covers a tiny amount of what's in there, and frustrating because there's little online discourse about any of it to refer to.

Is this earth shattering in itself? Probably not. But it's certainly better than you can find from "an online forum like Reddit or a book/paper published more than a year ago".

It's also far easier to find information through LLMs for another niche I play in, certificate authorities, than it is on the open web. Try to look up anything to do with certificates and you'll just get guided to endless "Why my certificate not work? Browser says no?" discussions, because actually creating certificate authorites is again very niche , not widely discussed and a lot of what you need to leverage in any given language isn't well documented because "you'll never need to do this". Well actually...

You're the one making the claim, and also claiming that you have specific examples to back up your point. So maybe you should share them, and we can see if your assertions hold up.

I guarantee that you could not find a proof using brute force, even with literally trillions of dollars of compute. You're betraying a lack of understanding of the complexities involved, unless you're using a definition of "brute force" which is essentially equivalent to "intelligent search."

> That is different from spending 10s of millions to do a brute-force search of lots of different proofs which can be verified by LEAN automatically

It’s kinda like the project where someone generated a lot of midi files for all possible pop riffs to copyright them. If you have a start and endpoint, rules for navigation, and unlimited computes, you can brute force anything. The nice stuff is when you made exploration efficient.

>That is different from spending 10s of millions to do a brute-force search of lots of different proofs which can be verified by LEAN automatically.

That is not remotely how the vast majority of the proofs have worked. You don't know what you're talking about, but that seems par the course for these discussions.