For those who are out of the loop, here is a TLDR of the situation:
Joint statement by the Lean and Mathlib communities, endorsed by the Lean FRO and the Mathlib Initiative.
On November 20, 2026, Levent Alpöge released a proof of the Riemann Hypothesis as a Lean library comprising 17 million lines of code.
We have full confidence that Alpöge’s proof of the Riemann Hypothesis is correct, and we congratulate him on this achievement.
Source: https://leanprover-community.github.io/blog/posts/riemann-hypothesis/


I’m no expert at all …
But I think the secret sauce that hasn’t quite made it into mainstream understanding is this:
It’s not one parrot any more. It’s a whole bunch working together with different assigned roles and prompts feeding one output into another as input and so on. Not unlike how society and the brain work.
Combine that with external systems that the models can use as a metric or a goal (such as “make this program return success”), and you have something not unlike evolution where the models and their inputs and outputs to each other are like mutations over generations, the external system like natural selection, and the models themselves like all the physics and biochemistry that is harnessed by life.