X · @togethercompute
· X / Twitter
We sat down with @tri_dao at ICML to ask where the next big architecture unlock comes from. His answer: There isn't one. It's kernels, inference stack…
We sat down with @tri_dao at ICML to ask where the next big architecture unlock comes from.His answer: There isn't one. It's kernels, inference stacks, and fleet optimization, layer by layer. And most of that work is happening in the open.