Achille and Soatto (Amazon/AWS) model AI reasoning agents as compute-capable stochastic dynamical systems and ask: can chain-of-thought reasoning solve any computable task? They show the key resource is inference time and that training data can exponentially reduce the time needed.
Found via https://www.amazon.science/blog/intelligence-isnt-about-para...
