My Mental Model of AI Broke on September 8
20 days ago
- The author's mental model of AI shifted on September 8, 2026, when a reported proof of a Millennium Prize problem (likely Navier-Stokes) was announced, suggesting AI may have produced new mathematical knowledge unknown to humans.
- This event is fundamentally different from previous AI benchmarks because it represents a genuine discovery beyond human knowledge, not just beating human performance on known tasks.
- The author previously viewed AI as a stochastic parrot or advanced search engine, but rapid progress from simple arithmetic to IMO gold medal and now a Millennium Prize problem shattered that belief.
- If the proof is correct, AI could move from helping with existing knowledge to producing new discoveries, accelerating a race among labs to demonstrate novel mathematics and science.
- The author warns of a potential crisis as society may struggle to adapt to AI-generated knowledge, impacting education, research, careers, and the concept of human intellectual uniqueness.
- The hope is that humanity can steer this transition toward improving human life rather than leaving people behind.