News for hackers. Stuff that matters.
Re:Chiplet ASIC supercomputers for LLMs like GPT-4 by kraken12 on Wednesday July 12, 2023 @03:02PM (#36700032)
Maybe they could do something like AMD's GPU memory stacking, that is good for scaling, and of course they are using many chips not one chip..

There are no comments on this comment yet.