alt Hacker NewsMaybe I'm wrong, but it looks like the authors did not actually have any LLMs write or verify any code for their experiments. Instead, their experiments consist of simulating the simplified Markov chain model itself. They simulated their simple Markov chain and checked if the theorem's predictions matched empirical statistics. This amounts to a test not of their model, but of basic Markov chain theory.
Did I misread or miss something?
Also, the mathematical content here is pretty thin. Their main theorem has nothing to do with LLMs directly. It's a theorem about a five-state Markov chain, and the proof follows from standard Markov chain theory.
For those reasons, the grandiose name "LLM-Verifier Convergence Theorem" does not sit well with me.