Which brings to the question of how the current hyperventilation around #llm, #ai and compute intensive algorithms and #datascience in general will reshape the #cpu/#gpu stalemate, the lack of an obvious winner in the post #mooreslaw quest for more computing power.
While the true utility of all this new software remains to be seen, it does feel that it will induce a Darwinian evolution/extinction phase, not unlike previous episodes e.g., the rise and fall (and rise again?) of #risc
2/2