The Rise of OpenAI and the Importance of Safety

OpenAI hasn't really explained to us what data they trained it on, the amount of computing resources they use for it. Could it be possible that OpenAI are further down the road to AGI than we really discuss? It's a very challenging coordination problem where the incentives encourage racing rather than careful slowed down coordination. And so I'll actually give you a quite concrete example that brings the race to life.