OpenAI Bulk-Solves 722 Math Problems, Anthropic’s Cruelty Policy, and Elon’s Quadrillion-Dollar SpaceX | EP #302

AI has reportedly "bulk-solved" an entire academic field (math) in days, a pattern panelists expect to next hit biology, medicine, and encryption.

• Math 'cooked' in a week: OpenAI posted 722 math solutions to GitHub on Oct 6 from an unreleased model (nicknamed Bell/GPT-7), averaging ~3 hours compute per solve, covering 81% of major math results from the past three years; notably zero cryptography results were included.
• Anthropic bans AI cruelty: Starting Nov 12, Anthropic prohibits 'sustained and needless' abuse of Claude, letting it end abusive chats, citing uncertainty about Claude's 'moral status'; Elon Musk says RL training traces look like interrogation under duress.

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.