Not lengthy after the phrases “996” and “grindcore” entered the favored lexicon, individuals began telling me tales about what was taking place at startups in San Francisco, floor zero for the synthetic intelligence economic system. There was the one concerning the founder who hadn’t taken a weekend off in additional than six months. The girl who joked that she’d given up her social life to work at a prestigious AI firm. Or the staff who had began taking their footwear off within the workplace as a result of, nicely, in case you had been going to be there for no less than 12 hours a day, six days per week, wouldn’t you somewhat be carrying slippers?
“In case you go to a restaurant on a Sunday, everyone seems to be working,” says Sanju Lokuhitige, the co-founder of Mythril, a pre-seed-stage AI startup, who moved to San Francisco in November to be nearer to the motion. Lokuhitige says he works seven days per week, 12 hours a day, minus a couple of fastidiously chosen social occasions every week the place he can community with different individuals at startups. “Typically I’m coding the entire day,” he says. “I don’t have work-life stability.”
One other startup worker, who got here to San Francisco to work for an early-stage AI firm, confirmed me dismal images from his workplace: a two-bedroom condo within the Dogpatch, a neighborhood standard with tech employees. His startup’s founders stay and work on this condo – from 9am till as late as 3am, breaking solely to DoorDash meals or to sleep, and leaving the constructing solely to take cigarette breaks. The worker (who requested to not use his title, since he nonetheless works for this firm) described the state of affairs as “horrendous”. “I’d heard about 996, however these guys don’t even do 996,” he says. “They’re working 16-hour days.”
Startups have by no means been significantly glamorous. After I began reporting on the business a decade in the past, individuals had been cashing in on the brand new cellular app economic system, and coders had been chugging Soylent to remain at their desks longer. Startups then, too, had been outlined by hustle tradition, high-octane power and the pursuit of development in any respect prices – concepts that, to some extent, have remained within the bloodstream of the business.
However within the final 12 months, because the magic mud of synthetic intelligence has settled in San Francisco, the vibe amongst tech employees does appear completely different. The joy a few new epoch in tech – and all the cash that comes with it – is now tempered with anxieties concerning the business, and the economic system. Some employees are going all in on AI whereas additionally questioning whether or not all that AI is sweet for the world. Others are successfully coaching machines to do their jobs higher than they’ll. And lots of the similar employees who’re racing to construct the long run are actually questioning if the long run they’re constructing has a spot for them in it.
The remainder of us could also be ambiently conscious of those anxieties, however they’re already tangible and keenly felt contained in the tech business. Even the largest tech firms, as soon as identified for coddling workers with on-site massages and barber outlets, have scaled again perks as they’ve escalated the expectations of employees. Mark Zuckerberg and Elon Musk have every been candid about their predictions that AI will substitute some junior and mid-level engineers at their firms, and have respectively known as for his or her workforces to be extra “environment friendly” and “extraordinarily onerous core” as waves of layoffs set workers on-edge. Tech firms laid off a few quarter of 1,000,000 employees all over the world in 2025, in accordance with a report printed by RationalFX. In lots of these layoffs, AI was cited as a predominant issue, even when the total purpose for layoffs is usually extra advanced.
“In case you had been a software program engineer 5 years in the past, you would type of write your ticket,” says Mike Robbins, an govt coach who has labored with firms like Google, Microsoft, Salesforce and Airbnb. Now, the stability of energy has shifted away from tech employees, lots of whom are left feeling anxious about their work efficiency. “When firms turn into much less scared about dropping workers, then they could be a little extra forthright when it comes to what they need and be slightly extra demanding.”
Robbins, who wrote the e-book Deliver Your Complete Self to Work, was requested to talk to firms and their leaders about subjects like worker burnout, wellbeing and belonging – high priorities within the years throughout and shortly after the pandemic. “Fairly frankly, we’ve stopped speaking about all that,” he says. Now, firm leaders need recommendation on subjects like change, disruption and uncertainty within the office.
These themes – change, disruption and uncertainty – are every a part of the gasoline that has pushed tech employees to place in additional hours, at the next depth. Funding in synthetic intelligence firms reached document highs in 2025, but employees are feeling shortage in methods they haven’t earlier than.
“It’s positively one thing that’s on everybody’s thoughts,” says Kyle Finken, a software program engineer at Mintlify, which makes an AI instrument for builders. “I believe lots of people are involved like, ‘Oh, am I going to have a job in three years?’”
Regardless of his fears, Finken, like many different startup workers I spoke to, feels energized by the “extraordinary innovation” taking place in synthetic intelligence and believes that there’ll nonetheless be loads of jobs for software program engineers sooner or later, even when these jobs look completely different from the pure coding roles of in the present day. He and different tech employees characterised the present second as a very artistic and productive time in tech, the place persons are devoting further hours to work not as a result of their employers demand it however out of real curiosity within the new instruments and capabilities. For instance, Garry Tan, the top of the well-known startup accelerator Y Combinator, just lately bragged that he “stayed up 19 hours” taking part in round with Claude Code.
Even those that felt excited concerning the tempo of change acknowledged that AI was quickly augmenting their work, in ways in which may have unsure outcomes for the roles of the long run. “That is positively not an period of complacency,” says Finken.
One purpose for working so many hours is to maintain up with instruments and expertise which are altering practically daily. In case you take the weekend off, you possibly can miss a significant growth, which makes it tougher to maintain up with what opponents are doing. One more reason is to have one thing to indicate future employers, particularly as extra junior-level jobs are changed by AI.
“Nobody hires junior builders any extra,” says Lokuhitige, the Mythril co-founder. Touchdown a job now requires “doing one thing cool”, he says, like constructing a brand new product or fixing an issue that will get acknowledged as helpful by bigger firms. Job postings for entry-level tech jobs have dropped by a 3rd since 2022, in accordance with Certainly’s Hiring Lab, whereas job postings requiring no less than 5 years of expertise have risen. In case you’re not grinding at a startup, you’re lacking the prerequisite to get employed sooner or later.
What this implies for the remainder of us
Whereas economists are torn about whether or not AI will substitute most jobs or simply change them, they appear aligned in the concept that AI has already reshaped a substantial amount of entry-level work and can proceed to take action. A paper printed by Stanford researchers in November discovered “substantial declines in employment for early-career employees” in industries uncovered to AI and steered that areas the place change is already occurring could possibly be like a “canary within the coalmine” for the remainder of the economic system. The Anthropic CEO, Dario Amodei, has steered AI may remove about half of all entry-level jobs in white-collar industries inside the subsequent 5 years.
The top of the Worldwide Financial Fund just lately predicted that 60% of jobs in superior economies will likely be eradicated or remodeled by synthetic intelligence, “like a tsunami hitting the labour market”. In San Francisco, you possibly can already see the early indicators, as Uber drivers compete with self-driving Waymos, and baristas are changed by robotic espresso bars. Skilled enterprise companies that assist the tech business have additionally been negatively affected by the layoffs. The strain to grind within the tech world could possibly be an early sign – a harbinger for what many different industries will really feel quickly.
Robbins, the chief coach, says that firms as soon as appeared to Silicon Valley as a mannequin of how they need to function, right down to emulating insurance policies like limitless trip days or adopting perks like free lunch within the workplace.
“There was an idealization of tech and Silicon Valley for a very long time throughout the enterprise world. A few of that has modified,” he says. “Now, individuals aren’t asking me to inform them what’s happening within the Valley in order that they’ll undertake it, the identical manner they had been a decade in the past.”
Relatively than a mannequin of how we must always all work, the tech business could also be a premonition for the anxiousness and makes an attempt to compensate which are coming for all of us.
