Steven Adler, ex-OpenAI safety researcher, analyzes major AI news, policy, and research - and what's needed for an actually exciting AI future.