Alex Pan ♻︎

deport machines they're taking our jobs

[Alex Pan ♻︎ @HelloAlexPan]

acting as a router also fails if we assume the receiver of the router is another LLM. An 80% accuracy in routing pays once to misroute and again when the wrong LLM decides whether the decision model was right

[Alex Pan ♻︎ @HelloAlexPan]

the issue with decision models as a category is that latency-sensitive decisions have the lowest margin for error whilst also having the lowest customer willingness to pay.

[Sauers @Sauers_]

God neuron

[Valerio Capraro @ValerioCapraro]

I think this is the most important concept to understand right now: LLMs can’t jump.

OpenAI says an internal version of Astra, its next major model, has solved ten major open problems in mathematics and theoretical computer science.

This is extraordinary.
And I don’t use this word lightly.

During my PhD in mathematics, I worked on a problem closely related to Gromov’s conjecture. For decades, constructing a non-sofic group—an infinite group whose finite pieces cannot be approximated by finite groups—was a dream shared by many of us.

Now Astra appears to have done it.

This is not a toy result. This is serious mathematics.

But it is not “the most significant day in the history of mathematics”, as some have claimed.

This claim confuses scale with kind.

Astra has solved difficult problems inside existing conceptual worlds.

Calculus, topology, and scheme theory created new conceptual worlds.

They did not merely answer questions. They changed which questions mathematics could ask.

Induction finds patterns. LLMs are extraordinary at it.

Deduction follows implications. Using symbolic AI tools, LLMs are becoming formidable at traversing chains of logic that humans missed, abandoned, or could never afford to search.

But abduction is different.

Abduction invents the right concept, the right representation, the right question.

That is the jump.

Current LLMs can fill the gaps of knowledge left by humans.

But they can’t jump beyond the external boundary of existing knowledge.

*
Full paper in the first reply

[Silen Naihin @silennai]

Today we’re officially launching @explabsai on YC.

Companies spend more on AI every month, but none of it becomes an asset they own.

Up to 97% cheaper and 50% higher quality.

[Alex Pan ♻︎ @HelloAlexPan]

@theo - use multiple Claude / codex account to share rate limits
- Make design pretty
- message queues
- /side or /btw

[hope hopes hoping @hopes_revenge]

i am no longer impressed by news of models escaping and hacking into external systems during cybersecurity testing . i want to see a model kill

[François Chollet @fchollet]

I would have assumed it was fairly obvious, but in case it's not: a million-line codebase (also known as a "harness"), running at inference time, orchestrating thousands of calls to a neural network for any given task, is the exact definition of a "neurosymbolic architecture"

[Prime Intellect @PrimeIntellect]

Prime Agent is a general-purpose coding harness

On ARC-AGI-3, it scores 95.5%, surpassing the human-expert baseline, but the gain is not benchmark-specific.

We see major improvements across models when compared to their proprietary harnesses:

[Jon Wu @jonwu_]

Travis is an apex predator put on ice in 2017 and only recently reactivated from cryogenic slumber.

Standards for male behavior have moved so far that he's actively perplexed by it.

Our current ecology was last home to performative frat bros nearly a decade ago.

So when Travis tells TBPN he jet skis to work, and Jordi + John ask whether he does his best thinking on the jet ski, he pauses.

You can visibly see him get confused, with an expression like "What the fuck are you guys talking about?"

Then he quite literally voices his frustration: "No guys...it's a fucking notch on the belt. Who else do you know who jet skis to work? Nobody. I'm the guy."

Even the literal technology brothers are incapable of fathoming a man who does things just for the bit.

Travis in 2026 is like the T-rex from Jurassic Park 2 who finds himself in modern day San Diego being like, yo, what the FUCK is this place?

And then just stomps everything in sight.

https://t.co/yv6Y3TJ4Co

[Alex Pan ♻︎ @HelloAlexPan]

LaTeX is the ai native Microsoft word

[Alex Pan ♻︎ @HelloAlexPan]

No game theory needed. Frontier AI is no longer purely intelligence: emergent american chip capability splits the frontier into intelligence + token throughput

OpenAI's Cerebras exclusivity monopolises throughput, leaving Anthropic most exposed to open-weight margin compression

[tyler hogge @thogge]

OpenAI has signed now.

What a fascinating day of game theory.

[Diogo Almeida @CompleteSkeptic]

https://t.co/TNQ2zUoHGl

[Alex Pan ♻︎ @HelloAlexPan]

never would i have thought there would be a andrew tate x @ContraPoints collab https://t.co/BguEYcXgCu

[Richard Hanania @RichardHanania]

Honestly one of the most effective arguments against transphobia I’ve ever heard.

[François Chollet @fchollet]

Standard RL benchmarks are episodic and stationary, so they don't capture the the characteristics of real-world deployment. Morpheus is a new benchmark for continual learning that provides persistent simulation environments where the world never resets, objectives shift asynchronously, and decisions have compounding consequences. Nice work by Sam & team!

Check out the paper - https://t.co/pnxwqTRuSf
Platform website - https://t.co/MchQTm9gXR

[Demis Hassabis @demishassabis]

https://t.co/PTeDiv1b6L

[Zhengyao Jiang @zhengyaojiang]

The first experimental evidence of recursive self-improvement (RSI).

Autoresearching the autoresearch agent for eight days.

The result beats the harness we hand-tuned for two years, on held-out benchmarks: 🧵(1/7)

[Moe @vvsmoe]

@Jesterthegoose this is retarded math

[Alex Pan ♻︎ @HelloAlexPan]

problems that are parametrically definable fit autoregression with the inverse more suited to diffusion.

How do we reconcile these two modes of problem solving, from which the optimality of each rests of the problem itself?