Founders Are Pulling All-Nighters Just to Keep Up With Their AI Agents

As our glossary explains, an AI agent isn't just a chatbot that answers questions — it's AI that can search the web, create files, and even make bookings on its own. Because it plans and executes multiple steps independently, tasks can finish or hit a wall while the founder is asleep, and each snag requires a human to sign off before the agent can move to the next step. On top of that, the models powering these agents are improving on a timescale of months, forcing founders to keep rewriting how they work.

A day in the life of four founders

Founder	Company	Work pattern	Notes
Aditya Sharma, 27	Keel	Pulls all-nighters, sleeps starting around 6 a.m.	Started using a smartwatch to check on agents remotely
Peter Pezaris, 56	Proxon	7:30 a.m. to 2 a.m.	Estimates roughly 30x faster output; projects over $1M in annual revenue by October
Abby Grills, 33	Riveter	10-hour days, one day off a week	Only two co-founders; leans on agents instead of hiring
Ajay Kalia, 43	Alt	5:30 a.m. to 9–10 p.m.	Manages agents via Apple Watch, which buzzes for approval every 10 minutes

Pezaris, who has spent more than 30 years in tech and successfully sold four companies, had promised his wife he'd retire this year. As a young founder, he was hospitalized twice from overwork, and once collapsed at the office after a month of sleeping just an hour a night. Four months ago, he started his fifth company, Proxon, anyway. With a team of six engineers, the company now runs an estimated 30 times faster than it would without agents. But there's a catch: as agents pile on more work, the human workload grows too — faster onboarding, for instance, means more customer support requests coming in.

Abby Grills and Cody Watters, co-founders of the data collection platform Riveter, still run the company as a two-person team. They keep thinking about hiring, but every time the question comes up, they find a way to hand the new workload to an agent instead — so hiring keeps getting pushed back. Grills went through Y Combinator's accelerator program in 2024, where she watched a founder from her cohort get hospitalized after struggling to eat solid food and surviving on meal replacements. That experience is why she now makes a point of actually resting when she takes time off. Y Combinator itself says it has long encouraged founders to take care of their bodies.

Ajay Kalia, who runs Alt out of Boston, said he sensed AI was about to upend everything back in February 2023, while working at Spotify's AI group, when colleagues with PhDs and 20 years of experience started saying things like "my PhD is worthless now." He left Spotify in 2024 to start Alt, and recently began controlling his agents through his Apple Watch — which now buzzes every 10 minutes asking for approval, meaning he has to check his wrist even in the middle of a run.

The FOMO of "missing a week's worth of progress"

Pezaris compared working with agents to being on drugs — he's never actually used drugs, he says, but he imagines this is what it feels like. Underneath that feeling is a kind of FOMO: the sense that every moment he's not working is a week's worth of progress lost. Sharma admitted he's gained weight over the past few months and finds it harder than before to spend time with people outside of work hours. Within Kalia's network of founders, he says, nobody talks openly about rest or burnout, but plenty of people seem to be running on fumes. Everyone quietly agrees it's not sustainable — but nobody's found a reason to stop.

The broader shift toward handing work to agents

This connects to something we covered earlier: One in Five US Workers Now Delegates Tasks to AI Instead of Colleagues.

That survey, conducted by Epoch AI and Ipsos among 1,106 U.S. workers last July, found that 20% said they were handing off work once done by colleagues or outside contractors to AI. The rate varied widely by task — from 25% for record-keeping up to 57% for systems and software design. Taken together with this WSJ reporting, the picture is clear: founders are offloading more and more work to agents, but the human burden of monitoring, approving, and handling exceptions isn't shrinking at all.

Editor's take

What this reporting really shows is that AI agents haven't turned out to be a labor-saving tool so much as a new workforce that needs supervising. We're nowhere near the point where a person just gives instructions and collects finished results. Right now, every time an agent gets stuck, a human has to step in immediately — so instead of cutting a founder's working hours, agents have stretched them into a 24-hour on-call shift.

The contrast with the old startup "9-9-6" culture (9 a.m. to 9 p.m., six days a week) makes this clearer. Back then, the limit on output was how many hours a person could physically work. Now, agents finish tasks faster than humans ever could, so the new bottleneck is founders failing to keep up with the agents' pace. Pezaris's estimate of "30 times faster" isn't an exaggeration — it's better read as a sign that as agents simultaneously handle onboarding, coding, and monitoring, the volume of approvals and reviews landing back on humans has exploded right along with it.

There's a lesson here for smaller Korean startups and teams considering agent adoption too: adding more agents doesn't mean you need fewer people — it creates a new kind of waiting time, where someone has to be available to make judgment calls whenever an agent gets stuck. Before following Grills's approach of leaning on agents instead of hiring, it's worth first measuring how many hours a day actually go into approvals and exception-handling. Skip that step, and you risk ending up with a founder who's effectively on call around the clock.

Expect to see more signs of this fatigue in the weeks ahead. It's likely we'll first see smartwatch apps for remotely controlling agents, or "agent manager"-style products designed to handle overnight approvals. What's actually needed isn't a way to slow agents down — it's a way to safely defer decisions without requiring a human to jump in immediately.