OpenAI's Sam Altman says that AI's benefits outweigh its risks, drawing a sharp contrast with Anthropic's safety push.