X · @soumithchintala
· X / Twitter
RT Horace He: In modern ML accelerators, FLOPS have absolutely exploded. Often though, the bottleneck is not FLOPS but memory bandwidth. Similarly, mo…
RT Horace HeIn modern ML accelerators, FLOPS have absolutely exploded. Often though, the bottleneck is not FLOPS but memory bandwidth. Similarly, model intelligence has exploded, causing the bottleneck to be humanAI bandwidth. At Thinky, we think that it’s important to solve this. 1/4Thinking Machines: People talk,