George Dover spent six years as a software engineer at Intuit Mailchimp in Portland, Oregon. By the time he was laid off in late 2024, his daily work had fundamentally changed. Instead of writing code, he spent his days using AI to generate code, then evaluating it for errors and bugs.
"The quality has to be rigorously tested," Dover said, describing his new reality as a software engineer who no longer wrote software — he audited machines that did.
After the layoff, Dover took a job that could not have been more different from his previous life: substitute kindergarten teacher. It was a world away from debugging code and attending sprint reviews, but it paid the bills while he searched for his next role in tech.
The job hunt was grueling. He sent out 400 applications over months, navigating a tech hiring market that had been fundamentally reshaped by the very AI tools he'd been using at work. Indeed tech job postings had dropped 36 percent from 2020 to 2025, and the unemployment rate for computer science graduates had risen to 7 percent.
Dover's boss at Mailchimp had warned him before the layoff to use AI more. "I am actively trying to keep my axe sharp," he said, referring to his efforts to maintain his engineering skills despite the industry's rapid shift toward AI-generated code. He described his future outlook as "dark."
Eventually, Dover landed an AI-oriented software engineering job — a role that barely existed in its current form a few years ago. But his journey from writing code, to reviewing AI code, to teaching five-year-olds, and back to tech, captures the disorientation that has become routine for an entire generation of software engineers.
He is part of a larger wave. Since ChatGPT's release in 2022, more than 600,000 US tech workers have lost their jobs. Computer science program enrollment fell 8.1 percent in 2025-2026, signaling that even students are reconsidering the career path that once promised the highest salaries in the American job market.