When AI Solves Your Life’s Work

The episode examines OpenAI's disputed revenue figures, The Information's survey on AI adoption, Anthropic's new Claude features and behavior policy, and then spends most of its time on OpenAI's release of hundreds of mathematics proofs. For CIOs, it signals fast-changing vendor economics, shifting model preferences, and the possibility that frontier AI can start disrupting knowledge work at field scale.

Key points

• OpenAI told investors annualized revenue was about $50B, not the widely reported $68B; the gap came from net versus gross accounting and short-window extrapolation.
• Markets treated the revenue report like an earnings miss, and AI-linked semiconductor stocks sold off sharply.
• The Information found 35% of respondents saying AI returns are multiples of spend, while only 8% called AI net negative.
• OpenAI regained share over Claude and Gemini; 72% of respondents now use open-weight models regularly.
• OpenAI's math release reportedly solved 90 of the top 500 open problems, raising verification and knowledge-management challenges for research teams.