Perhaps it could go wrong because he uses AI robots to generate responses to issues on claude code that are also generated by AI robots? Just bots talking to each other like moltbook. It shows a level of AI maximalism that is absurd, concerning, and funny. But probably par for the course for someone working at Anthropic. I can imagine being surrounded by people doing similarly foolish things only encourages the foolery.