“OpenAI's new multi-agent reasoning team will pursue a different approach than has historically been pursued by the multi-agent research community.”