OpenAI Codex developer says agent swarms waste tokens with no quality gain
Eric Provencher, a Codex developer at OpenAI, warned on X that large parallel agent swarms incur a "coordination tax"—burning tokens on redundant verification without improving output quality. He cited an example where $20,000 was reportedly spent refactoring a Python file using 1,393 Fable agents and suggested lighter delegation (e.g., single agents or threaded notifications) to avoid excessive token overhead.