Symbolic learning works

Some observations about the "total" automation of software engineering: 1. There are more software engineers now than in 2025, and demand i...

Argon has a similar performance profile on ARC-AGI-3 as Astra. Google is back at the frontier. https://twitter.com/arcprize/status/211076828...

I truly believe conversational AI will drive the biggest educational leap of our generation. 1-on-1 tutoring is the gold standard for educat...

You can't solve real-world physical tasks through autoregressive sampling alone. The physical world is entirely unforgiving to hallucination...

String theory is what happens when an intelligent system is starved of sensory input and starts hallucinating. Intelligence needs grounding....

The ideal outcome for science isn't a couple of labs using proprietary models and millions of dollars to crack trophy problems, it's to have...

If an agent saves an engineer 2 hours to implement a feature but collectively costs the team 3 hours in alignment and review, the ROI is neg...

Congrats on the release! Looks like a fantastic model. https://twitter.com/sundarpichai/status/2110038898009407584

If you explained agriculture to folks in SF, you'd blow so many minds. It's... self-replicating, recursively self-improving nanotech for the...

I've been having a lot of fun trying a wide range of LLM post-training experiments using KerasHub + JAX + TPUs. Might open-source some of th...

How AI shifted from a transductive paradigm (token-by-token answer completion) to an inductive paradigm (on-the-fly reasoning chain synthesi...

The real world abounds with recursively self-improving systems, but one in particular deserves attention: Science, modeled as a system (perh...

Right now, using AI agents still feels very single-player. They do your work, but the results are trapped in isolated chat histories and not...

We're reaching really good ARC-AGI-2 scores on Kaggle now! https://twitter.com/arcprize/status/2108578092558250148

Most measures of AI progress closely fit an exponential over the 2023-2026 period. Meanwhile AI capex over the same period has been slightly...

Humans provide a fun comparison point here. In the 80s, the "cognitive consequences of programming" was a major psychological research focu...

What if the jagged frontier is mainly math + code (which you can push arbitrarily far with RLVR), and everything else starts to plateau beca...

Arguing that "consciousness arises from a computational substrate, and since AI models are computation, they are likely conscious," is as me...

The notion that current AI models are sentient and can suffer, combined with the foolish idea that suffering can be mathematically quantifie...