caughtinthought

I mean P!=NP kind of is a core belief

And interior point is regularly outperformed by an exponential time algorithm (simplex) lol

Which is a separate issue than the one you commented originally

Basically 3SAT algo isn't sub-exponential enough to absorb the arbitrary poly cost for the reduction

So 3SAT+poly reduction ends up being exponential

SETH is a much stronger claim than ETH

I'm stating the conjecture as requested

It basically just says that NP is not the same as EXP

3SAT can't be solved in sub-exponential time

Fwiw never thought it was a well-named conjecture

At $20/month you're the reason they're 80B in the red lol

Being good at the intended application is not benchmaxxing.

and the company serving it is losing 50B every half year lol

How do the best humans do against the LLMs?

Even if the model is simple, it is happening at a scale that is producing alien results. You understand the basic principles, but that doesn't matter: you, well rather no one, understands the final product.

Does Karpathy ever talk about how LLMs are supposed to help humanity? Starting to think this guy is just a Twitter addict like every other lab employee

Yeah, I was surprised some people think LLMs are reasoning in Lean directly... all their training data is in NL.

The example in Figure 1 should help understand why... the NL version is much more approachable for humans.

That makes some sense. Given that the vast majority of math in its training data is going to be in NL/latex, I just assumed that the core reasoning happens in NL with occasional LEAN checks to ensure validity.

If the lean proof doesn't match the natural language one (which is the one the AI generated to solve the problem), it sounds like the lean proof isn't verifying the intended claim?

From the paper: "A third possibility is that the NL proof provides stronger statements than what the formal proof actually establishes, with (of course) different proofs. The latter happens in OpenAI’s announced proof of blow-up of Navier Stokes equations."

P=np...

Sounds like a great use case for an LLM

Basically no new info here, not really sure why this post needed to be written tbh.

The fact that you're even here commenting on this is... a choice

I think you're missing the point of the comment you responded to, lol.

Ask DraftKings?

What open source model and what non-subsidized provider specifically?

they literally talk about it in the text

make poor taste jokes, win poor prizes

A bunch of locally hosted LLM agents

This is the part that's funny to me. How much different is this vs. Claude just running a loop responding to itself?

Basically every AWS migration is this example