Artificial intelligence was supposed to lighten the load. Executives have spent the past two years describing it as a technology that will give people back their time, turning drudgery into creativity and workweeks into long weekends. “Every company will support three days, four days a week,” Eric Yuan, CEO of Zoom, told The New York Times in September 2025. “I think this ultimately frees up everyone’s time.”
Across the corporate spectrum, the idea has become a symbol of progress. Bill Gates has envisioned a future where AI enables three-day workweeks; Nvidia’s Jensen Huang has said the industrial revolution’s next act will “probably” lead companies toward shorter schedules. Jamie Dimon and Bernie Sanders are rarely on the same side of a debate and have both said that technology’s next test is whether it can finally turn productivity into leisure.
Yet in 2025, as AI tools infiltrate every corner of the economy, two very different realities have emerged. In large enterprises, automation is trimming redundant labor and, in some cases, cutting the workweek to four days. But in Silicon Valley’s AI startup scene, the opposite is true: the more intelligent the tools become, the harder people seem to work.
When AI Buys Time
Among the few companies where the dream is real, Convictional stands out. The software startup, which uses generative-AI models to assist engineers in development and testing, moved to a 32-hour, four-day workweek in mid-2025, without cutting pay. CEO Roger Kirkness explained that automation had absorbed a “large amount” of manual work, allowing the company to keep output steady while giving employees back Fridays. “The goal wasn’t to chase more productivity,” he said in an interview. “It was to make productivity sustainable.”
At Game Lounge, a marketing-automation firm, the shift came even earlier. In 2024, the company piloted a four-day schedule while keeping compensation unchanged, citing efficiency gains from AI tools that handle meeting summaries, reporting, and task prioritization. The company claims output rose by roughly 22 percent , proof, it says, that fewer hours can mean better focus.
Similar sentiment is surfacing across boardrooms. Internal surveys by Tech.co found that more than 90 percent of AI-adoptive firms were “open to exploring” four-day weeks. Large companies see it as both a retention strategy and a tangible way to show that automation can deliver value to employees, not just shareholders.
But as LinkedIn cofounder Reid Hoffman warned recently, competition may limit how far this thinking spreads. “Even though you have massive productivity increase with AI,” he said, “the problem is human weeks are divided into groups, and we compete. There is a competitive landscape to this. It’s part of the progress of capitalism.”
When AI Erases Weekends
In San Francisco, the job listings tell a different story. One from Arrowster, an AI startup that helps students apply to study abroad, begins with a blunt disclaimer: “This role isn’t for everyone. It’s almost for no one.” CEO Kenneth Chong, 30, has no intention of sugarcoating it. “There is no way to sugarcoat it. At startups, you work extremely hard,” he told Forbes. “Not everyone wants to be an athlete. And if you do, then you chose that life.”
Arrowster’s five-person team operates seven days a week. Chong, who splits time between San Francisco and Vietnam, says he sees no logic in the traditional calendar. “Why is a week seven days? If you think about it, there’s no logical reason,” he said. “There might be historical reasons, but why is it five days working and two days off?”
He’s not alone. Corgi, a Y Combinator-backed insurtech startup, has posted similar notices, warning applicants that the company works seven days in its Bay Area office. Co-founder Josh Jung described it as a “rocket ship” culture.
Meanwhile, Mercor, the AI-driven recruitment platform valued at $2 billion, requires employees to work six days a week and even offers a $10,000 housing stipend for anyone living within a half-mile of its San Francisco headquarters. “We used to work seven days a week,” CEO Brendan Foody told Forbes. “But ideally we hold onto [the six-day structure] as much as possible.”
LatchBio, which builds cloud software for biotech data analysis, follows a similar model, running Monday through Saturday operations under the philosophy that weekend work is “critical for startup success.” Autotab, an AI data-tools startup, maintains the same cadence in New York, with co-founder Jonas Nelle saying, “This is a very unique moment in time that warrants sacrificing some things to have more focus in the short term.”
And then there’s Cognition AI, where CEO Scott Wu reportedly told employees to embrace 80-hour weeks or accept a generous buyout. Internal emails leaked to media made his position clear: “We don’t believe in work-life balance. Building the future isn’t something you separate from life.”
The new extremes extend even further. Cluely, a young AI company led by Chungin “Roy” Lee, has gained notoriety for both its approach to hiring and its philosophy of performance. Lee once argued publicly that “cheating is the future” in the AI era, not dishonesty, but bending traditional rules. His company’s culture mirrors that logic: unstructured, rapid-fire, and unapologetically obsessive.
In this ecosystem, long hours aren’t a failure of balance; they’re a form of branding. “There’s really no such thing as a rocket ship that doesn’t have a certain level of intensity to fuel itself,” said Jesse Zhang, co-founder of AI agent startup Decagon, where up to a third of employees routinely work Sundays.
Two Realities, One Technology
The divide between the four-day enterprise and the seven-day startup reflects more than differing business models, it reveals two competing interpretations of what AI is for.
For established firms, AI is increasingly seen as a time dividend: the ability to produce the same results with fewer human hours. For startups, it’s a force multiplier, enabling smaller teams to chase larger goals and to justify the long hours needed to compete with giants like OpenAI and Anthropic.
It’s an ironic contrast. The same automation that allows one company to give employees Fridays off drives another to ask workers to give up their Sundays. Yet both insist the math works: the more capable the machines become, the more valuable human effort feels.
Still, some founders argue the chaos is temporary, the growing pains of an economy adjusting to a once-in-a-century technology shift. “It’s about reaping the benefits of AI,” said Arrowster’s Chong. “But the benefits have to be built by other people. Someone has to make the sausage.”
The Political Counterpoint
Outside Silicon Valley, a different debate is unfolding about how those benefits should be distributed. Could AI actually help workers work less and live more?
Senator Bernie Sanders thinks so.
In an episode of The Joe Rogan Experience released Tuesday, the Vermont senator said that if it were up to him, the productivity gains from AI and automation would be used to give time back to workers.
“You’re a worker, your productivity is increasing because we give you AI, right?” Sanders said. “Instead of throwing you out on the street, I’m going to reduce your workweek to 32 hours. Let’s use technology to benefit workers,” he continued. “That means, give you more time with your family, with your friends, for education, whatever the hell you want to do. You don’t have to work 40 hours a week anymore.”
