Latent Space: Why There Is No 'AlphaFold for Materials' — AI for Materials Discovery with Heather Kulik

Latent Space·medium signal

Latent Space interviews MIT's Heather Kulik on why a decade of AI for materials science hasn't produced an AlphaFold-like breakthrough. Unlike protein folding which had a clear evaluation metric (CASP), materials discovery lacks standardized benchmarks and faces combinatorial complexity orders of magnitude larger than biology. The episode covers what's actually working in AI-assisted materials research, where the field is stuck, and lessons for AI practitioners about domain-specific challenges that benchmark-chasing misses.