Raw model intelligence is argued to be becoming less of a constraint than UX, speed, and habitual work patterns. The source is skeptical that society or most companies are actually capped by intelligence today, and expects near-term optimization effort to shift toward "good enough," cheaper, faster, locally-run models rather than the frontier.
This reframes where product and engineering effort should go: once a model is smart enough, gains come from interface design and habit formation, not from chasing the next frontier model.