Tech leaders say AI means less work - their staff say they work up to 90 hours a week
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI formally urged companies in 2025 to test four-day work weeks with no pay cut, but a former technical employee said the firm never trialed it internally, describing weekend work, frequent "crisis meetings," and "super cut-throat" performance reviews that saw colleagues suddenly let go
- At Anthropic — whose CEO Dario Amodei has claimed Claude can work seven hours without a break and warned of broader AI-driven job losses — sprints can stretch beyond 90 hours in a seven-day period, per workers who spoke to the BBC; neither OpenAI nor Anthropic responded to comment requests
- Meta workers said they were abruptly "drafted" onto AI teams this year without a choice — "You can't say no – or if you do, you have to quit" — while the company laid off roughly 1 in 10 of its employees, according to one current and one former staffer
- Former Google employee Amin Shali left in May, saying the shift of processing units and memory storage to AI projects was causing engineering failures and forcing him to work through the night; his sleep and health improved after leaving, and a Google spokesperson declined to comment
- A UC Berkeley study tracked hundreds of tech workers over eight months and found AI use led to faster pace, broader task scope, and extended hours, while MIT scholar Neil Thompson said saved time gets "sucked up" by new work emerging from the tools
Why it matters: For the workers building frontier AI tools, the productivity dividend leaders keep selling is being captured by employers, not employees. UC Berkeley's eight-month finding that saved time gets absorbed by new tasks means the four-day-work-week pitch risks becoming a rationale for higher output expectations rather than shorter hours — and roughly 1 in 10 Meta staff already paid that bill in layoffs.
Ask SkimNews
