very surprising that fifteen years of hardcore computer vision research contributed ~nothing toward AGI except better optimizers we still don't have models that get smarter when we give them eyes

→ View original post on X — @jxmnop