The claim is stark, and it lands with a thud: artificial intelligence is set to wipe out many roles traditionally held by humanities graduates, while hands-on trades and technical vocations become the new safe harbor. In a moment when white‑collar confidence feels shaky, Palantir’s chief executive offers a provocation and a prediction: the future belongs to people who can actually make, fix, and install things.
“Software is getting good enough to do repetitive writing, summarizing, and research,” he says, “but it can’t crawl into an attic, sweat a pipe, or rewire a panel.” The message is unsettling for office workers—and oddly optimistic for anyone with a toolbox.
White-collar tremors, blue-collar tailwinds
AI’s rapid competence at language, drafting, and analysis compresses value in roles that trade in text, templates, and predictable reasoning. The pressure is fiercest where work can be standardized and where errors are cheap.
Humanities‑adjacent jobs—copywriting, basic research, entry‑level communications, grant drafting—sit squarely in AI’s strike zone. “We’re not talking about a few tasks,” the CEO warns. “We’re talking about the core of how a lot of knowledge work gets done.”
By contrast, fields that merge physical coordination with situational judgment—electricians, machinists, HVAC techs, line installers—face rising demand and slower automation. The bottlenecks aren’t purely technical; they’re regulatory, logistical, and deeply human.
Why the trades look resilient
Robots and models still struggle with messy, unstructured reality. The world of buildings, roads, hospitals, and factories throws endless edge cases at any automated system. Apprenticeships transmit tacit knowledge—the smell, sound, and feel of a problem—that textbooks can’t fully encode.
“People who can fix real things will have their pick of work,” he argues. “We are going to have a surplus of software that writes software, and a shortage of people who can keep society running.”
Expect steady demand where tasks are:
- Physically situated and variable (rooftops, crawl spaces, roadside repairs)
- Regulated and licensed (electrical codes, welding standards, medical tech)
- Customer‑facing and trust-sensitive (home services, onsite diagnostics)
- Integrated across systems (mechanical‑electrical‑plumbing coordination, factory retrofits)
The education reset that follows
If the analysis holds, the talent pipeline must shift. Four‑year degrees will remain powerful, but the blanket default looks dated. Short‑cycle programs, paid apprenticeships, and stackable credentials gain prestige.
High schools could re‑center shop, robotics, and field labs; colleges could pair philosophy with cybersecurity, history with data, ethics with engineering. The goal is hybrids: people who can think critically and turn a wrench with confidence.
“Societies that respect craft do better in crisis,” the CEO notes. “We need fewer performances of competence and more proofs of competence.”
How companies can prepare now
Enterprises should map which roles are automatable, which are augmentable, and which are structurally embodied. Invest in tools that make technicians super‑productive: AR guides, digital twins, predictive maintenance, and mobile workflows.
Recruit for “head‑and‑hand” talent: installers who can query data, machinists who can tweak code, nurses who can orchestrate AI triage. Reward outcomes, not just output volume, and track real‑world uptime as a metric.
For workers, the advice is blunt and practical: learn one tool of automation, one tool of measurement, and one tool of repair. Put yourself close to the bottleneck—the step that customers most value and machines can’t yet do.
Not a simple victory lap for vocational work
There are caveats. Robotics and embodied AI are advancing, and what looks safe today may feel less so tomorrow. Some trades will fragment into AI‑assisted modules, pushing wages down unless policy and standards keep pace.
Cultural inertia also bites. Societies often glamorize screens and desk jobs while underselling the dignity of skilled labor. If compensation, benefits, and career ladders lag, the pipe of new talent will remain thin.
Where the humanities still matter
None of this says humanistic thinking is obsolete. It says the easy, commoditized tasks of language are getting cheaper. The premium moves to judgment, narrative framing, ethics, public trust, and human‑machine governance.
Pair rhetoric with risk, literature with policy, anthropology with operations. The winners won’t just produce content; they will shape consequences, architect systems, and explain trade‑offs to communities.
“AI can draft a brief,” he concedes. “It can’t own the stakes, absorb the blame, or build consensus in a tense room.”
Choosing the future we want
If AI hollows out the middle of knowledge work, leaders must rebuild a middle in the real economy. That means sponsoring apprentices, modernizing vocational schools, reforming licensing that blocks mobility, and celebrating excellence in the trades.
The paradox is clear: as software eats more text, society needs more touch. The safest careers may be those with a little grease under the nails—and a lot of judgment above the eyes.
