Are we dismissing AI spend before the 6x lands?

Critics are judging models trained on last-gen hardware. There's a 6x wave of compute already allocated - and it's just starting to produce results.