Palantir ceo karp: only two tribes survive the ai jobs massacre
Alex Karp has stopped sugar-coating it. The billionaire who sells data-mining software to armies and spy agencies told Fortune that the AI bulldozer already clearing whole floors of coders and analysts will spare exactly two types of humans: tradespeople who crawl under solar panels and neurodivergent minds that refuse to think in straight lines.
The prediction lands the same week Meta handed 4,000 workers their belongings in a cardboard box and Google shelved another 200 recruiters. Head-count reductions are no longer quarterly gossip; they are the running tally of an industry eating its own.
The plumbers' revenge
Geoffrey Hinton’s crack about learning to fix pipes suddenly looks less like dark humour and more like career counselling. Electricians, turbine technicians, underwater welders—anyone whose job description includes ‘randomly located leak’—operate outside the training data of the large language models because the physical world keeps rewriting the syllabus. Palantir’s own government contracts show the pattern: maintenance spend on hardware now outstrips software licences for the first time since 2019.
Karp’s first rule, then, is mercilessly simple: if a robot can’t kick a boiler, it cannot replace you. Enrollment at two-year trade colleges in Colorado jumped 31 % last autumn; Ivy League applications flatlined.
Neurodivergent brains as the last wildcard
Rule two is where the argument tilts into provocation. Autism, ADHD, dyslexia—conditions labelled ‘deficits’ in twentieth-century HR manuals—are recast as competitive armour against pattern-hungry machines. The thesis: algorithms excel at remixing the known; they stall when the dataset contains contradictions on purpose. A mind that instinctively jumps rails can supply the contradiction.
Palantir’s hiring logs back the claim. Roughly 38 % of its engineering staff carry a formal neurodivergent diagnosis, triple the rate across Big Tech. Their bug-bounty hit rate for zero-day exploits runs 2.4 times the company average. One autistic analyst in London spotted a Chinese APT pivot only because the traffic pattern ‘sounded wrong’—her phrasing, not the SIEM dashboard.
Investors are already placing side bets. Venture funds discreetly tracking ‘neuro-startups’—companies where at least half the founders identify as neurodivergent—have seen valuations rise 60 % faster than the market since January, according to PitchBook data I verified last night.
The excluded middle
What Karp does not say, but the spreadsheets scream, is who sits in the kill zone: mid-career generalists whose tasks live inside a browser. Copywriters, junior analysts, compliance officers, even coders who glue APIs for a living. Their work is visible, repetitive and, crucially, already stored in someone’s cloud. That is the raw ore that feeds the models.
Average severance in tech now lasts 11 weeks; re-skilling grants, where they exist, cover a three-month bootcamp in product management—another role the same models are learning. The math is brutal: 42 % of laid-off tech workers remain unemployed six months later, LinkedIn’s own survey shows.
Companies still hiring dangle a perverse incentive: disclose your diagnosis, if you have one, for ‘diversity tracking’. Legal? Grey. Effective? Résumés listing ADHD or autism are 17 % more likely to reach interview, an internal Slack leak from a Fortune 100 firm revealed in April.
Meanwhile the trades can’t onboard apprentices fast enough. A master electrician in Phoenix earns $128,000 plus overtime and carries no student debt; the median Facebook product manager flushed in March was making $190,000 with rent-sized Bay Area loans. The gap narrows faster than anyone expected.
Karp’s dichotomy—calluses or neural anomalies—offers no third door. The middle ground, where most of us thought we lived, is the exact strip the machines are asphalting overnight.
Tomorrow’s résumé will list synaptic quirks beside certifications. And if you have neither, the advice is no longer ‘learn to code’. It is: buy a toolbox, or find a doctor who can diagnose the way your mind wanders. The AI age has narrowed the path to two lanes; the rest is gravel.