New data shows US entry-level software jobs have fallen 23% since ChatGPT launched, while senior roles have grown. Kobeissi pulls together labor data that puts hard numbers on a debate that's mostly been anecdotal until now: is AI actually replacing junior workers? US software developer employment for the 22-25 age group has fallen 23% since November 2022, the month ChatGPT launched. The 26-30 cohort is down a smaller but still real 5%. Compare that to developers aged 41-49, whose employment has risen 18% over the same stretch, a 41 percentage point gap between the youngest and one of the most experienced groups in the same profession. The pattern holds even after controlling for firm-specific shocks: workers in the most AI-exposed early-career roles saw a 16% employment decline relative to less-exposed peers and more experienced colleagues, while less AI-exposed fields and senior staff kept growing. For investors, this cuts two ways. It's a real-world data point that large language models are already substituting for the kind of routine coding work junior engineers used to do, bullish evidence for the AI productivity thesis that's driving valuations across the sector. But it's also a warning about the consumer and macro side: an entire cohort of young, credentialed workers locked out of their first jobs is the kind of labor-market scar that shows up later in consumer spending, delinquencies, and political backlash against the technology driving it.
Why it matters · Hard labor-market numbers, not anecdotes, showing AI substituting for junior coders is either the clearest proof yet of the AI productivity story or an early warning sign for consumer spending and political backlash.
Worth asking · Is a 23% drop in entry-level coding jobs since ChatGPT proof AI is already delivering on its productivity promise, or a warning sign for the next generation of workers?