CHATGPT RISE PUTS DEVELOPER HIRING UNDER PRESSURE

A new Federal Reserve working paper is putting numbers behind one of the most sensitive questions in the technology labor market: whether generative AI has already slowed hiring for software developers.

- 

The study, written by economists Leland D. Crane and Paul E. Soto, finds that employment growth in programming-intensive occupations in the United States decelerated sharply after ChatGPT launched in November 2022, with the slowdown concentrated in jobs most exposed to large language models. The paper does not claim that ChatGPT directly eliminated half a million existing jobs. The more precise finding is that developer employment continued to grow, but much more slowly than it did before late 2022. Based on the researchers’ counterfactual estimates, the gap between actual employment and the path that programmer employment was likely on before ChatGPT represents roughly 500,000 “missing” jobs over three years. That is a measure of foregone growth, not mass layoffs.

That distinction is crucial. The study is not saying every one of those positions was cut by AI. It is saying the market for coders appears to have changed around the same time large language models became widely visible to executives, managers, and workers. The authors identify November 2022 as the natural breakpoint because ChatGPT's launch made the productivity implications of AI coding tools far more obvious, even though GitHub Copilot and earlier models were already available before then.

WHAT THE FED STUDY MEASURED

Crane and Soto focused on computer programming-intensive occupations because coding is one of the clearest areas where large language models can assist or substitute for human work. AI tools can now write boilerplate code, suggest functions, debug errors, translate between programming languages, summarize documentation, and help less experienced workers complete technical tasks faster.