Jason Fried on AI work culture: products aren't getting better at the rate the development process is improving
Jul 22, 2026 with Jason Fried
Key Points
- Jason Fried argues that AI is accelerating software development speed faster than product quality improves, creating a mismatch where users cannot absorb constant changes into their workflows.
- 37signals founder observes that AI tools are producing burnout rather than the promised free time, with teams managing parallel workstreams without clear outcomes.
- Fried frames the core problem as one of discipline: faster tools raise production capacity but don't answer what should actually be built.
Summary
Read full transcript →Products aren't getting better as fast as development is
Jason Fried's sharpest observation has nothing to do with the individual burning out in the viral screenshot. It's structural: AI is dramatically accelerating how fast software gets built, but customers aren't seeing the benefit at anywhere near the same rate.
"I'm not seeing products get better at the rate that the development process is getting better," Fried says. The gap isn't just about quality — it's about absorption. Users don't want products changing constantly. They want to settle into something, understand how it works, and build it into their routines. More output doesn't fix that mismatch; it can make it worse.
“I do find it ironic that AI is what it is, yet everyone seems to be working harder and harder. I'm not seeing products get better at the rate that the development process is getting better. People are doing a lot of stuff, and yet people actually don't want their products to change rapidly either. Just because you can build more of it doesn't mean it makes it a better product — can make it a worse product — and you're seeing that all over the place right now.”
The implication is pointed. Teams shipping faster with AI assistance may actually be degrading their products by moving too fast for users to follow. Fried frames it as a curation problem: "You still have to decide what gets through the slit."
On the burnout question, Fried finds the irony hard to miss — a technology that promised more free time is producing exactly the opposite. At 37signals, he says he can sense the weight of it: parallel agents, multiple workstreams running simultaneously, and a creeping question of "to what end?" He stops short of condemning the tools — "the tech is incredible, obviously" — but argues that technology which exhausts people isn't serving them well, however impressive it is under the hood.
The viral screenshot itself — about the San Francisco engineer who locked himself away for months to become his company's top AI user — gets a drier read. "Top user" likely means most tokens consumed, not best outcomes delivered. Whether the three-month AI bender translated into real product value is, as Fried notes, unknown.
The through-line is a discipline argument: access to better tools raises the ceiling on what teams can produce, but it doesn't resolve what they should produce. Speed is an input. Customers judge outputs.
Every deal, every interview. 5 minutes.
TBPN Digest delivers summaries of the latest fundraises, interviews and tech news from TBPN, every weekday.