How far is ZKP from compute verification in high-stakes environments?

A baseline overhead measurement and the hash-problemThis is an exploration into the current state of ZKP for LLM inference to use in a high-stakes environment like international agreements.