It's not the job. It's the mindset.

In March, Palantir’s chief executive Alex Karp went on the tech show TBPN and gave Gen Z the bluntest career advice of the year. There are basically two ways to know you have a future, he said: “One, you have some vocational training. Or two, you’re neurodivergent.”

Fortune resurfaced the clip this month and it’s doing the rounds again. Read as a list of approved jobs, it’s billionaire provocation. Read as a description of how people think, it’s one of the more useful things anyone has said about surviving AI.

Because neither of Karp’s categories is really a job. One is a way of working. The other is a way of seeing. That’s the angle we care about here: not which job title is safe, but which mindset is.

The tradie mindset

The case for trades is real. Nvidia’s Jensen Huang told Davos in January that the data-centre boom is the “largest infrastructure build-out in human history”, with six-figure work in it for electricians and plumbers. Fortune, writing about Carhartt’s chief executive Linda Hubbard, puts skilled trades at 32% more pay than other jobs that don’t need a degree.

In Australia, employers filled only 55 in every 100 advertised trades and technician jobs in the first three months of 2026, the lowest of any kind of work. And that shortage isn’t unrelated to AI. Line up all eight kinds of work by how hard they are to hire for, and by how much of each one Jobs and Skills Australia reckons AI could automate, and they fall along a slope.

Hard to hire for, hard to automate

Up: how many of every 100 job ads employers filled, January to March 2026. Across: how exposed the work is to being automated by AI, on Jobs and Skills Australia's 0 to 100 scale, averaged across the jobs in each group. Tap or hover a dot for its numbers.

Kind of work	Ads filled (%)	AI could automate (out of 100)
Clerical and administrative workers	83.8	60
Community and personal service workers	70.8	30
Labourers	56.3	20
Machinery operators and drivers	60.8	28
Managers	79.1	40
Professionals	66.5	39
Sales workers	64.9	48
Technicians and trades workers	55.1	27

Ads filled: Jobs and Skills Australia, Occupation Shortage Report, March 2026. Automation: Jobs and Skills Australia, Our Gen AI Transition (2025), an unweighted average of the occupations in each group. © Commonwealth of Australia, CC BY 4.0.

That is a pattern, not a cause. Trades are hard to hire for because the training takes years, not because a chatbot can’t wire a house, and sales sits a little off the line. But the shape is hard to miss: the work that’s easiest to hire for is the work a machine could most easily take over.

But look at why trades hold up. It isn’t the toolbelt. It’s how a good tradie works:

• They walk into a mess nobody planned for and work it out on the spot.
• They own the result. If it leaks, it’s their name on it.
• Many run their own business, so there’s nobody further up to pass the problem to.

That matters, because the job itself won’t stay safe forever. Robotics firms already argue general-purpose robots will move into the trades once the labour shortage makes it pay; one, RobotLAB, expects the first commercially viable robot plumber by the end of this decade. Whether or not they’re right about the date, the clock is ticking on the job. The mindset (diagnose, adapt, own it) travels to whatever comes next.

Thinking differently

Karp’s second category is personal. He has long spoken about his dyslexia, and Palantir runs a dedicated Neurodivergent Fellowship. Gartner even predicts that one in five sales organisations in Fortune 500 companies will actively recruit neurodivergent people by 2027.

But listen to how Karp explains it and the diagnosis fades into the background. What he’s describing is a disposition: people who are “more of an artist”, who look at things from a different direction, take risks and build things nobody else would.

A diagnosis isn’t a career plan. Originality is. AI is superb at producing the most likely answer. The premium shifts to people who reliably don’t.

The other camp agrees more than it thinks

Karp is often set against tech leaders who defend a broad education. But they end up in the same place.

Microsoft’s chief scientist Jaime Teevan argues that what counts now are metacognitive skills: “flexibility, adaptability, experimentation, thinking critically.” Anthropic’s cofounder Daniela Amodei says the things that make us human will matter more, not less, and that her company looks for people who are great communicators, kind, and curious. And Huang reminds anxious workers that the purpose of a job and the tools you use to do it “are related, not the same.”

Different words, same idea. Nobody serious is saying a credential will save you. They’re all saying a way of thinking might.

Five mindsets that beat the bot

Pull the threads together and you get a short list. None of them is a job title. All of them can be learned.

1. Hands-on. Stay close to the real, messy world, where neither software nor robots are comfortable yet.
2. Contrarian. Look at the problem from the direction nobody else is facing. The obvious answer is the one AI already has.
3. Owner. Take responsibility for the outcome, not just the task. Machines don’t carry accountability; people do.
4. Adaptive. Separate who you are from the tools you use. The tools will change every year.
5. Human. Be the person others trust, understand and want to deal with. Empathy doesn’t automate.

Which ones do you already bring?

Tap the ones that sound like you, and watch them go up on the board. Nothing is saved or sent anywhere.

With JavaScript on, this is a five-button check with a little classroom whose pinboard lights up as you tick each mindset.

The more of these a person brings, the less it matters which job they’re in.

From both ends of a career

I’m about to turn 54. I’ve spent my career in digital product design, at Apple and now in banking, the kind of screen-based work AI is reshaping fastest. Many of the junior designers on my team are my daughter’s age.

She graduated in product design this year and is moving into fashion design: physical materials, taste, a customer who touches the thing. People might say she picked the safer lane. I’d say she picked a mindset: hands-on, a little contrarian, close to the people she designs for.

The same choice is open to me, and to every junior designer on my team. You can’t always change your job. You can change how you do it.

The bottom line

Karp named two groups. What he was really describing is one mindset: people who work close to reality and think in their own direction. Jobs will come and go, and the robots are coming for physical work too, just more slowly. The question isn’t which job is safe. It’s are you the kind of person who stays useful whatever the job becomes?

That’s what we’re building this site to help you answer. Find your job, and read what it’s made of.