Palantir’s karp says only two tribes will survive the ai jobs purge
The robots are already inside the building. While analysts still debate when artificial intelligence will ‘disrupt’ the labour market, lay-off spreadsheets at Google, Meta and a dozen Fortune 500s tell a simpler story: the future arrived early and it’s wearing a pink slip.
Alex Karp, the unshowy billionaire who runs Palantir, watches the bloodletting from his corner office in Denver and shrugs. In a recent Fortune interview he distilled the next decade into two lanes: ‘You either have a trade, or you’re neurodivergent. Everyone else is overhead.’
Why electricians keep winning
The first lane is easy to picture. Solar-panel installers crawling across Arizona roofs. Elevator mechanics in Manhattan shafts. Plumbers, the punch-line of every tech conference joke, now bill more per hour than the product managers who once mocked them. Geoffrey Hinton, the ‘godfather’ of deep learning, advised students to drop Python and pick up a pipe wrench. Karp simply monetised the gag: Palantir’s share price tripled after it started selling data analytics to the same blue-collar supply chains Silicon Valley promised to erase.
The numbers back him up. The U.S. Bureau of Labor Statistics projects 800,000 new openings for skilled trades by 2032 – jobs too site-specific, too finger-on-circuit for a cloud API to swallow. AI can read a wiring diagram; it can’t crawl through a crawlspace and rewire a 1950s bungalow while dodging a raccoon.
The hidden superpower of misfits
Lane two is stranger. Karp’s ‘neurodivergents’ – the ADHD coder who hyper-focuses for 36 hours, the autistic analyst who spots market noise no one else hears – are suddenly assets, not head-case HR files. Large-language models gorge on pattern; they choke on anomaly. Palantir’s government contracts rely on analysts who think diagonally, the ones the Pentagon calls ‘weapons-grade weird’. Inside those minds, Karp sees the last moat.
It’s a brutal filter. Roughly 15 percent of the population qualifies as neurodivergent. Add the trade cohort and you still haven’t covered a third of today’s workforce. The rest – spreadsheet jockeys, middle managers, content moderators – occupy the evaporating middle. AI doesn’t need their permission to replace them; it just needs cheaper electricity.
Companies feel the shift in real time. IBM has paused hiring for back-office roles it expects algorithms to absorb. Dropbox’s CEO told staff to prepare for an ‘AI-first’ structure, a euphemism that saved him from spelling out redundancies. Meanwhile, Palantir quietly recruits welders and data savants, the twin species Karp bet on five years ago.
The moral is crude, almost medieval. Learn to bend copper or bend reality. Everything in between is being automated before the press release hits the wire.
The lay-offs will keep coming, dressed up as ‘restructuring’ or ‘efficiency gains’. But the queues at vocational schools and the waiting lists for adult autism assessments tell the same story: the economy is already reallocating oxygen. If you’re neither handy nor wired differently, the clock isn’t ticking – it’s just flashing 00:00.