Product managers stare at screens, grinning through gritted teeth. Energized by AI’s raw power. Exhausted by its endless demands. Nikhyl Singhal, former Meta vice president of product, calls it ‘smiling exhaustion.’ ‘I’ve never seen an industry that’s more tired than they are now,’ he told Lenny’s Podcast. Everyone’s in a state of alert. Nothing’s constant.
Singhal sees the paradox everywhere in his community of senior leaders. AI tools like Claude let product managers prototype ideas, test hypotheses, coordinate teams—all faster than before. Tasks that dragged on for days now wrap in hours. Workers ship more. They feel productive. But stress mounts. Skills obsolete in months. ‘People, even if they’re doing well, they feel more stressed because they worry they’re either not keeping up, or they worry this industry is going to change, and essentially they’ll be roadkill along that way,’ Singhal said on the Business Insider-cited podcast. If you don’t love building stuff, you’re in trouble. The pace is relentless.
Software engineers echo the strain. Siddhant Khare, an ONA engineer building AI agent infrastructure, shipped more code last quarter than ever. Felt more drained, too. ‘AI fatigue is the kind of exhaustion that no amount of tooling or workflow optimization could fix,’ he wrote in his essay AI fatigue is real and nobody talks about it. Each task shrinks—design docs, scaffolding, tests. But days balloon. Six problems, not one. Context-switching kills. Humans tire. AI doesn’t.
Khare became a reviewer on an endless assembly line. Prompt. Wait. Judge. Fix. Repeat. AI code demands scrutiny; it’s confident but suspect. No shared history with colleagues. Every line a potential pitfall. ‘We used to call it an engineer, now it is like a reviewer,’ he noted in a Business Insider profile. Nondeterminism grinds worst. Prompts that nailed Monday flop Tuesday. No why. Just vigilance, always.
Simon Willison, Django co-creator, pushes further. Fires up four AI agents parallel. Tackles separate problems. By 11 a.m., wiped out. ‘Using coding agents well is taking every inch of my 25 years of experience as a software engineer, and it’s mentally exhausting,’ he said, per Business Insider. Others lose sleep, letting agents grind overnight. Intensity surprises. AI saves time. Demands more brain.
Studies confirm it. A Harvard Business Review report, ‘When Using AI Leads to Brain Fry,’ hit in March. Surveyed 1,488 U.S. workers. Fourteen percent report mental fog, headaches, slow decisions—’AI brain fry.’ Highest in marketing, HR, ops, engineering. One to two tools boost output. Three? Diminishes. Four-plus? Drops. Cognitive load from overseeing agents overwhelms, per lead author Matthew Kropp of Boston Consulting Group in Business Insider. ‘It’s a very specific effect… high cognitive load.’
Silicon Valley amps the frenzy. AI startups import China’s 996 grind—9 a.m. to 9 p.m., six days. Sarah, 28, at a hot SF firm, beyond burned out. Breakfast to bedtime work. Personal list spirals. Wife handles home. ‘We basically never hang out,’ she said. ‘What we’re all thinking is, ‘When does this level out?’’ OpenAI teams pull it. One job post spells 70-hour weeks. Equity lures. Bodies break, per PC Mag. Chronic stress warps brains—exec function, emotions. Inflammation rises. Burnout locks in.
Layoffs pile on. Over 85,000 tech jobs cut in 2026’s first quarter. Developers hit hardest. AI coding tools spike output 40-55%. Ten do fifteen’s work. Headcounts shrink. Entry roles vanish. Hybrids rise—those directing AI, not just coding. Singhal predicts mass shedding: 30,000 product staff out, 8,000 AI-first rehired in 12-24 months.
X buzzes with it. Developers lament productivity’s curse. ‘AI made peoples lives so efficient. Most people who worked 8-9 hours a day now work 14-16 hrs,’ posted @asaio87. Sergio Pereira warns execution-only engineers: pressure shifts to owning end-to-end. @Hesamation shares Khare’s essay: ‘You’re not imagining it… the industry is aggressively pretending [AI fatigue] doesn’t exist.’
But cracks show adaptation. Willison hoards techniques in GitHub repos, remixing for agents. Khare built Distill for deterministic context. Singhal urges ‘moments of joy’—building what excites. Companies eye limits: fewer agents, enforced breaks. Some push four-day weeks; studies show gains in health, satisfaction.
The boom powers on. Tech talent powers it. Grinning. Grinding. Headed for a wall—or reinvention. Older workers see labor compression: more from fewer, cheaper. Young ones chase the upgrade. Reality splits them. AI didn’t kill jobs. It redefined them. Survivors direct the machines. Thrive on the chaos. Or become roadkill.
WebProNews is an iEntry Publication