Zuckerberg's Retraining Promise Masks Permanent Wage Collapse
Mark Zuckerberg is telling workers not to worry about AI, that companies will always need people, that jobs will change and skills will adapt and the economy will find equilibrium.
This is what executives say when they are about to displace thousands of people and want to feel good about it. Meta laid off 10,650 employees in November 2022—Zuckerberg called it necessary.
Now he's on the circuit explaining why AI job losses are overblown, why anxiety about automation is misplaced, why historical precedent shows workers figure things out. He's not wrong that workers adapt—he's wrong about what happens to them when they do.
In 2008, major banks received the same diagnosis. The financial system needed restructuring, inefficient operations would be eliminated, and workers would retrain and move into new sectors. This was the language of creative destruction, and it dominated every boardroom conversation about the recession's aftermath. What actually happened lives in wage data.
Workers displaced during 2008 who found new employment took jobs paying 15 to 20 percent less than their previous positions. A manufacturing worker became a retail worker—a financial analyst became a logistics coordinator. The jobs existed, but the wage floors didn't. A decade later, earnings for that cohort still hadn't recovered. Retraining programs helped some people, but they didn't restore the economic position that automation had erased.
Zuckerberg is not claiming Meta will retrain its laid-off workers for new AI-era positions. He's making a broader cultural argument while his own company proves its narrower commitment. Efficiency comes first. Worker transition comes later, if at all.
The difference now might be scale and speed. AI doesn't displace workers in one sector at one moment. It moves through the entire economy simultaneously, collapsing the time workers have to find equivalent work. That's not adaptation. That's a floor giving way while the ladder is still being built.
