[abalaji]

> the scientific insights that got us to systems like GPT-4 and o3 were hard-won, but will take us very far.

Does anyone know if there are well established scaling laws for reasoning models similar to chinchilla scaling. (i.e. is the above claim valid?)