Viewing profile — empath75

You could imagine different layers on different chips, though, i think...

Those providers are not going to end up being price competitive with the big labs unless they are running at the scale of the big labs.

This is one of those headlines where I simultaneously think "cool" and now have a new source of low-grade background anxiety about. It can have a seat next to false vacuum collapse…

> In a world where open-weight models dominate, the proprietary frontier models of Anthropic and OpenAI will find it virtually impossible to charge monopolistic pricing. It'll neve…

Of course, the price of tokens going down makes it cheaper. You can now afford to throw millions of tokens at solving a problem that would have been impossible for AI to solve at a…

LLM's reason and think in the way that boats "swim" and planes "fly". It is an analogy. But nevertheless, planes do take you where you want to go, and so do LLMs. I think we are in…

That very much does not describe what happened. Someone found the bug and used it to disprove the Collatz conjecture as a demonstration of the bug. Nobody ever claimed it as an LLM…

A lot of the AI aesthetic of UX _currently_ is a function of the defaults in the skills and tools that ai agents are equipped with. That's distinct from the "AI look" for image gen…

I think the Balogun red-card and just inconsistency in general with red card suspensions was the only real blight on the tournament and that's fixable with some rules changes. The …

I think sports gambling should be legal _and regulated_, just like a lot of other vices, and one of those regulations _absolutely has to be_ a complete firewall between sports bett…

I've been using TLA with claude code at work and it roughly takes twice as long, but it's already _fast_ to get claude to produce code and this prevents a lot of rework.

This is the equivalent of a security researcher having a virus escape a sandbox. It's negligent, it's not criminal.

A lot of people thought that OpenAI was making this up, and I hope if you believed that, that you recalibrate your opinions of what LLM's are capable of. Working with Fable and Opu…

You can get _better_ that SOTA AI performance by burning the weights directly to silicon, the problem is you're locked into that model

I'm currently writing such a language myself in pure Lean, based on adjoint logic -- as well as graded modes and effects. I started by just trying to formally verify a Rust-like bo…

I had claude spend about 2 hours friday modeling one of our kubernetes operators in TLA, and it found 8 bugs, including 2 that could cause data loss, and one of the other bugs was …

I didn’t. Claude wrote all the proofs, I just validated that it was sorry-free, didn’t have any extra axioms other than mathlib and that it proved what i wanted it to prove (there’…

I spent two weeks proving a number theory result myself in lean just to see what I could do. (It’s something about composing polynomials with themselves and what other polynomials …

It does mean that this will have a drift problem if it's just trained on the idiosyncrasies of model fine tuning. That's fine! But it is something to be aware of.

I'm basically of the opinion that if agents can't write your programming language, that's a problem with your programming language.

I think there's a germ of a good idea here, but really this needs to be data that is presented in a way that encourages human thought and interpretation and not something claude pr…

The problem right now with something like this is you're trying to nail jello to a wall. People haven't figured out what an agent is yet, and trying to crystalize what people happe…

There's so much wrong with this comment. First, middle class _by definition_ did not have large estates that earned incomes for them. Second, it's weird to throw in an "unfortunate…

People made fire for thousands of years before they understood what it was.

Almost every single statement in this sentence was wrong. It was founded as the Sacrum Imperium -- literally the Holy Empire. "Heiliges Römisches Reich" was a later German translat…