lukeplato

> Our models produced a natural language proof for Yang–Mills existence on September 9th, 2026, and then subsequently produced a proof for QCD on September 18th, 2026. Given the length and complexity of the work, we postponed release until the Lean formalization completed. We finished our formalization of Yang–Mills’ existence on October 16th, 2026, which led us to develop the collaboration described here

absolutely insane

incredible that ideas like this are still being invented

Many comments seem to be misinterpreting this, but all this seems to be suggesting is that the AI companies should not be treating mathematical milestones as benchmark targets but need to release the cumulative findings in a human interpretable and digestible way that enriches and develops the field. See Thurston quote on Tao's banner:

> We are not trying to meet some abstract production quota of definitions, theorems and proofs. The measure of our success is whether what we do enables people to understand and think more clearly and effectively about mathematics.

related paper on the long-horizon AI research system used: https://arxiv.org/pdf/2608.11195

they're probably referring to CASP13/14

I think an unrestricted model with sufficient inference time compute could create unprecedented catastrophes.

Not sure why you would be opposed to whistleblower protections

hopefully nuclear power plants

Michael Edward Johnson has an interesting theory called vasocomputation with the core hypothesis:

> vasomuscular tension stabilizes local neural patterns. A sustained thought is a pattern of vascular clenching that reduces dynamic range in nearby neurons. The thought (congealed pattern) persists until the muscle relaxes.

https://opentheory.net/2023/07/principles-of-vasocomputation...

https://x.com/johnsonmxe/status/1863603206649208983

I don't see why a mixture of experts couldn't be distilled into a single model and unified latent space

reading this thread was a good reminder that being intelligent but closed off to alternative hypotheses is the same thing as being ignorant

I remember reading that for k-d trees to be able to split on k dimensions the dataset needs to be > 2^k, which becomes unwieldy pretty quickly

This could make for an interesting UI for exploring clusters in data. I only wish K-d trees could handle higher dimensions

interatomic potential is a promising scale and there's been some interesting recent developments using equivariant graph neural networks: https://www.nature.com/articles/s41467-022-29939-5 https://arxiv.org/pdf/2206.07697