A.I. Isn’t Coming for Every White-Collar Job. At Least Not Yet.

In January, Perry Metzger used Codex, an A.I. code generator from OpenAI, to help build an online word processor. What would have taken him and a partner at least two months was finished in two days, though he warned that “you have to keep a close eye on what it is doing and make sure it doesn’t make mistakes, and create ways of testing the code.” Tools from OpenAI, Anthropic and others have rapidly changed how software gets written, shifting many experienced programmers into supervisory roles as A.I.

does more of the hands‑on work. The systems also attract people with no programming background, who can use plain English prompts to assemble apps and agents. Researchers at Carnegie Mellon found that A.I. generators can produce code more quickly but often introduce lower‑quality work and technical debt that slows projects over time.

codex, openai, anthropic, code generator, prompts, software development, programmers, technical debt, carnegie mellon, supervisory roles