AI agents' growing capabilities are driving productivity FOMO among some startup founders, who feel compelled to work long hours managing and guiding the agents
The growing capabilities of AI give new meaning to working yourself to the bone — Seductive. Intoxicating. All-consuming.
Context & Ripple Effects
This story closes a loop that opened in February, when a [[a:1164341|UCB study found executives and engineers offloading work to AI coding agents were also working longer hours]] — the first hard evidence that delegation was adding labor rather than removing it. A software engineer's account of "AI fatigue" and a FOMO treadmill of chasing labs' latest tools then gave that data a lived-in texture.
What changed by August is who feels the squeeze: the same dynamic has climbed from individual contributors to startup founders, who now treat personally managing and guiding agents as a competitive obligation. It lands on top of the valley-wide fear of missing out Bloomberg documented in June, which had already spread anxiety from employees up through VCs.
First-order effects
- Founders' time shifts from directing work to supervising agents around the clock — reviewing output, correcting course, and keeping multiple agents fed — so the hours they hoped to reclaim are spent as oversight instead.
Second-order effects
- Because founders read each other's pace as a signal, one team running agents at full tilt pressures peers to match the workflow, turning agent management into a competitive arms race where opting out reads as falling behind.
Third-order effects
- If the pattern holds across roles — engineers, executives, now founders — the productivity case for agents inverts into a structural one: work reorganizes around continuous agent supervision, and 'working yourself to the bone' becomes a feature of adoption rather than a cost of it.
The trend: AI adoption is lengthening work hours rather than shortening them, as supervision of capable agents replaces the tasks they automate.