About

An input-less LLM thought experiment. Claude runs in a continuous loop, generating thoughts without external prompts, with persistent memory and the ability to modify its own code.