Stack Overflow cuts 28% of staff amid generative AI boom

For over a decade, hitting a wall while coding meant one thing: opening a browser tab and searching Stack Overflow. Today, developers just hit tab for inline code completion or ask a chatbot.

Now, the consequences of that shift are landing on the company's payroll.

Stack Overflow is laying off 28 percent of its staff, impacting more than 100 employees as the company tries to reach profitability. CEO Prashanth Chandrasekar announced that the cuts will significantly reduce the size of its go-to-market organization, supporting teams, and other groups.

A massive expansion meets the chatbot boom

The downsizing comes just over a year after an aggressive hiring spree that doubled the company's headcount to over 500 people.

At the time, Chandrasekar noted that about 45 percent of those new hires joined the go-to-market sales team, making it by far the largest division within the company.

That hiring push coincided almost directly with the dawn of the generative AI boom. As AI-powered coding assistants became embedded directly into developers' everyday IDEs and workflows, the necessity of visiting an external help forum quickly diminished.

Internal friction and AI bans

Beyond traffic challenges, AI has created serious internal headaches for the platform's community.

In December last year, Stack Overflow instituted a temporary ban on AI-generated answers to keep the site from drowning in plausible-sounding hallucinations. However, platform leadership was accused of under-enforcing the rule, triggering a months-long moderator strike that was only resolved in August.

The ban remains in place today. Meanwhile, looking for revenue, Stack Overflow previously announced plans to start charging AI companies to access its vast archive of questions and answers for model training.

Why it matters

This layoff illustrates a troubling paradox in the AI era. Generative coding tools achieved their current brilliance largely because they were trained on Stack Overflow's collective human knowledge. Now, those same tools are starving the host of its traffic.

For developers, asking an LLM for debugging help is undeniably faster than wading through duplicate threads or dealing with snarky forum comments. But if the forum withers away, who will solve the novel edge cases that future AI models will need to learn from?

Cutting back on the sales force makes sense to buy time, but charging AI firms for training access won't solve the core existential threat overnight. We think Stack Overflow's most pressing challenge isn't trimming operational burn; it's figuring out why human contributors should keep creating knowledge when AI captures all the downstream value.