Comment by xyzzy123

Ok I guess the team with the largest LLM workload in the world and billions in funding won't understand how to optimise a chip for the exact workload they have and near future ones.