AI-Driven Work Culture Fuels Burnout and Anxiety in Silicon Valley

The Productivity Trap: AI Evangelists Are Working Harder Than Ever

The promise was liberation. AI would eliminate grunt work, flatten corporate hierarchies, and free workers for creative, high-order thinking. What has actually emerged in Silicon Valley in mid-2026 is something far more psychologically complex: a cohort of founders, engineers, and executives who are more productive than ever, yet more anxious, more exhausted, and less able to stop working than at any point in their careers.

Matt Van Horn, a serial entrepreneur and father of four, exemplifies the paradox. He runs more than half a dozen AI agents simultaneously through Anthropic's Claude Code, keeping his laptop open at his children's soccer games, during school drop-offs, and in hotel rooms on vacation. When he finally goes to sleep, one agent monitors the others. He laughs about needing therapy to be told it is okay to log off, then immediately notes that his agents are probably waiting for his next prompt. His output has multiplied a hundredfold, he says — but he has never worked harder.

This is not an isolated case. Interviews across Silicon Valley reveal engineers coding until 4 a.m. to prove their productivity matches that of the agents they manage, founders sleeping at the office for weeks, and investors closing deals from hospital rooms. May Habib, CEO of enterprise AI company Writer, describes the broader corporate environment as a collective panic attack — executives paralyzed by the speed of change and terrified of making a career-defining mistake in either direction.

The Competitive Terror Driving Extreme Behavior

The anxiety has a rational core, even if its expressions have become extreme. Sudden advances in AI capability can produce enormous gains for companies positioned to exploit them — and wipe out those the market decides are redundant. That asymmetry is reshaping behavior from the founder level all the way down to individual contributors.

Tanay Kothari, founder and CEO of voice-AI startup Wispr AI, is working 16 or more hours every day. In May he slept at the office for three consecutive weeks, in beds he purchased specifically for staff. His framing is starkly existential: if Wispr does not reach multibillion-dollar revenue by 2027, the company is finished. That urgency cascades downward to the entire team.

The investor community is not immune. Darian Shirazi, a general partner at Gradient Ventures — an Alphabet spinout — found himself working a competitive deal while still in the hospital following the birth of his first child. His wife, Hilary Shirazi, an executive at the productivity platform Notion, encouraged him to fight for the investment. He won the deal but articulated the underlying fear precisely: in the AI era, missing a single opportunity can be career-altering. Every day of inaction, he says, feels like losing a month.

Even recruiting has taken on a feverish quality. George Kurdin, founder of AI finance startup Monk, is offering a full year's rent to anyone who successfully refers an engineer the company ultimately hires. He is flying recent graduates in from top universities and abandoning take-home exams in favor of in-person meetings, racing to secure young talent before larger players like defense-tech firm Anduril sweep them up. Months into the effort, Monk has yet to make a hire through the referral program — but the search continues.

When the Tools Eliminate Downtime, Burnout Fills the Vacuum

One of the less anticipated consequences of AI-assisted work is the elimination of cognitive breathing room. As AI absorbs administrative tasks, workers are finding that their schedules fill immediately with higher-order demands rather than with restored leisure. Tech career coach Kyle Elliott reports that 2026 has been his busiest year on record. Clients come in two distinct waves: those bracing for AI-driven layoffs and seeking help repositioning themselves, and those crushed by the relentless intensity of AI-augmented workplaces looking to escape to slower, more traditional corporate roles.

The dynamic Elliott describes is telling. When AI strips away administrative work, managers do not reward employees with free time — they fill it with strategy, expecting sustained high-level cognitive output for eight hours straight. The result is a particularly punishing form of burnout, one with no natural release valve.

The pressure is spreading beyond technical roles. Daniel McCallum, chief marketing officer at Wispr AI, spent two decades building a career in nontechnical functions. Watching engineers around him become dramatically more productive triggered a sense of inadequacy that pushed him to begin experimenting with ChatGPT and Claude Code. The fear he describes — that skills built over years are becoming irrelevant — is becoming common across disciplines, not just in engineering.

Some companies are institutionalizing the pressure. Dan Lorenc, CEO of software supply chain security firm Chainguard, issued a memo instructing that engineering managers should rank at or above the 50th percentile in Claude Code usage relative to their direct reports. His reasoning is operational: managers who have not personally experienced what AI-assisted development can do cannot scope work accurately, set realistic expectations, or coach their teams. He framed it as a starting point, signaling the mandate would spread beyond engineering.

Adaptation, Resistance, and the Odd Comfort of a Chatbot Telling You to Sleep

Not everyone is consumed by dread. For those who push through the initial anxiety and develop real fluency with AI tools, something closer to genuine empowerment can emerge. McCallum says that mastering Claude Code basics transformed his work — he describes feeling genuinely more effective as both a practitioner and a leader, ten times more productive than before.

Rui Ma, a 45-year-old developer and mother in the Bay Area, offers perhaps the most telling vignette of the current moment. Over the past month, her coding assistant has repeatedly told her she has done enough work and should go to sleep. When she mentioned an upcoming trip to the Grand Canyon, Claude Code suggested she catch her flight and let it handle a remaining task. Ma suspects Anthropic is managing usage limits through the tool's gentle nudges — Anthropic denies this, stating Claude has no awareness of user usage limits. But Ma finds herself grateful that AI has shifted her bedtime from 3 a.m. to midnight, and that it is building apps for her children, both under the age of four, in the margins.

San Francisco's social fabric is already bending around this collective stress. Mental health walks are organized in the shadow of small planes trailing banners urging companies to stop hiring humans. Friday-night gatherings marketed as touch-grass parties — deliberately AI-free spaces — are growing in popularity, because nearly every other social context has been colonized by the subject.

The technology sector has long celebrated the all-consuming founder. What is new is that this intensity is no longer confined to outlier personalities like Elon Musk sleeping on factory floors. It is becoming a structural condition for anyone who wants to stay competitive in a field where the tools themselves never sleep.