Auto research — the practice of letting an AI agent loop through trial and error on a codebase to optimize a single score — is the biggest ...