> Every ALU operation is a trained neural network

So basically: ChatGPT, what's 2+2 ?
Too bad, because the idea of emulating a CPU on GPU is a pretty good one. You do like 10,000 threads and the GPU cycles through one instruction at a time, helping whatever threads make progress that it can.
@cjd @HN
you posted article this to trigger me, didn't you? xD

imagine the interesting things people could do with heavily parallelized architectures, but they choose to do the most uninteresting unimaginative crap possible.
Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.