Recent submissions

Comment · 27 days ago
Comment · 1 month ago
>OpenAI has 'cheated' with this in the past, claiming that AlphaZero [...]

That doesn't sound right

Comment · 2 months ago
The whole point of Lean is that you don't need to understand the entire proof to be sure that it's correct. You only need to understand the definition of the theorem being proven, and you need to trust that the relatively small core of Lean is correct.
Comment · 3 months ago
That doesn't really answer the question of how they're going to detect it at scale. Ban repositories with long comments?
Comment · 3 months ago
> A 3rd party hosting platform like fireworks.ai that hosts open weights models won't provide any of that at all.

Why couldn't they?

Comment · 4 months ago
So then bots will just intentionally fail the test?
Comment · 4 months ago
> Over treatment can be a real problem.

Indeed, but having more data might be able to solve that? The whole problem seems to be that benign conditions sometimes look scary because we're currently not able to predict well enough whether it's something that will eventually cause problems.

Comment · 4 months ago
It's equally signaling that other US-based labs can't provide reliable access to their closed-weight models.
Comment · 6 months ago
Verifying that every step in a (potentially long) proof is sound can of course be much, much harder than verifying that a definition is correct. That's kind of the whole point.
Comment · 1 year ago
By "online wallet" they were likely referring to the Bybit website being the wallet of those customers that held their coins there rather than keeping them in their own private wallets, and not whether the hack involved a hot wallet or a cold wallet. Calling it a custodial wallet would have been more accurate.
Comment · 1 year ago
> It would be similar to if I claimed that an LLM is an expert doctor, but in my data I've filtered out all of the times it gave incorrect medical advice.

Computationally it's trivial to detect illegal moves, so it's nothing like filtering out incorrect medical advice.

Comment · 2 years ago
> in the absence of other celestial bodies the satellite would be in a stable orbit

Presumably entering such an orbit is only possible due to forces from other celestial bodies in the first place, since otherwise if you reversed time it would spontaneously leave its orbit. In other words, the act of the earth "capturing" the object is ultimately performed by external forces?

Comment · 3 years ago
They said nothing about not printing any non-primes.
Comment · 3 years ago
VanMoof bikes are known for breaking quickly. Calling them "so much better" than Cowboy based on a single bad experience doesn't seem totally reasonable.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton