Bold Predictions and the Future of Science

The Future of Creative Organizations ✏️ — with Dan Shipper

Topic 7 of 7

Published Mar 6, 2025 · Clip 44:31–48:00 (3:29)

Dan is most excited about o1/o3-class models getting tool use inside Cursor, since Claude gets lost in loops and o1 would let him code much faster. He offers a bold contrarian take: AI will fundamentally change science, because machine learning succeeds by finding thousands of small correlations rather than abstract causal theories, and he argues 'it's correlations all the way down'—correlation-based prediction will prove more effective than causation-seeking in many domains.