Career & Work
What Harari Got Right (and Wrong) About Work in 2026
Yuval Noah Harari predicted a lot of things about work and AI. Some of it held up. Some of it aged like milk. Here's the mid-decade audit.
In 2018, a bestselling historian predicted that people like you - knowledge workers, freelancers, anyone who types for a living - were on track to become what he called the “useless class.” Displaced by AI, with retraining too slow to keep up. It’s 2026 now, so his math is finally checkable.
Yuval Noah Harari made that call in 21 Lessons for the 21st Century, and the AI section in particular got passed around aggressively.
We’re halfway through the 2020s now. Enough time has passed to do an honest scorecard. Not “was he a genius?” or “was he wrong?” - both are dumb framings. The question worth asking is: as a freelancer, creator, or solo operator reading him today, what should you actually believe?
I’ll go through what held up, what didn’t, and what he missed entirely. No fanboying, no contrarian theater.
What he got right
The pace of disruption is unevenly distributed
Harari argued that AI and automation wouldn’t hit jobs uniformly - some categories would be gutted while others barely moved. This held up better than people give him credit for.
Translation, basic copywriting, junior legal research, customer support, mid-tier illustration - these are not what they were five years ago. Meanwhile, plumbers, therapists, electricians, and senior strategists are doing fine. The hollowing-out is happening in the middle of the cognitive labor market, exactly where he pointed.
If you’re a generalist content creator, you’ve felt this. The bar moved. Work that paid in 2020 doesn’t pay now, not because it’s worse, but because the supply curve shifted.
Reinvention is now a recurring expense
His point that you’d need to reinvent yourself multiple times in a career, not just once, has held up brutally. The half-life of a stable skill stack is shorter than it was. People who treated their 2019 toolkit as permanent got caught flat-footed.
I’ve written about this elsewhere - career pivoting without burning your life down leans on similar logic. The era of “pick a profession at 22, retire from it at 65” is over for most knowledge workers. Harari called it.
Storytelling skills compound
This one’s underappreciated. Harari spent a lot of time on the centrality of narrative - that humans coordinate around stories, not facts. In 2026, this looks more right than ever. The people thriving in AI-saturated markets aren’t the best at the tool. They’re the best at framing, packaging, and meaning-making.
The technical operator who can also tell a story about why the work matters? They’re up. The pure technician who relies on the work speaking for itself? Down.
What he got wrong
The “useless class” framing
This is the prediction that aged worst. Harari talked about whole categories of people becoming economically redundant, with no useful labor to contribute. It was provocative and got a lot of airtime.
The reality is messier and more interesting. The work changed shape, but humans found new ways to be economically useful - often in roles that didn’t exist five years ago. Prompt engineering, AI-assisted design, AI-output verification, hybrid creator-strategist roles. Whole categories of work materialized inside the disruption.
That doesn’t mean nobody got hurt. Plenty of people lost income they relied on. But the catastrophist framing - that entire populations would simply have nothing to do - has not held up. The labor market is more adaptive than he gave it credit for.
Timing on AI displacement
He significantly overestimated the speed at which AI would displace knowledge work, and underestimated how much augmentation would dominate over replacement.
In 2026, most knowledge workers using AI aren’t being replaced by it. They’re using it to do more, faster, sometimes worse, sometimes better. The “AI takes your job” headline keeps not arriving in the way the books predicted. What we have instead is a much weirder dynamic where the same person, with AI, displaces three people who don’t use it.
I’ve written about this in AI Won’t Replace You - But Someone Using AI Might. That post is the working theory I’d put more weight on now than the harder displacement framing.
Over-reliance on macro trends
Harari is a historian. He thinks in centuries. That’s powerful for some kinds of analysis and useless for predicting what a freelancer should do next quarter. A lot of what he wrote treats individuals as essentially powerless against the larger forces - and while that makes for compelling reading, it’s terrible advice for an actual operator.
The freelancers and solo workers who’ve done well in the last five years didn’t sit around waiting for the macro forecast to resolve. They moved, adjusted, repositioned, and kept moving. Macro thinking is interesting. Micro action is what pays the bills.
What he barely touched
Two things Harari mostly missed that turned out to matter a lot:
The premium on trust. AI made content cheap and abundant. That made trust expensive and scarce. The people who built reputational capital before the AI boom are in a different league now than the ones who didn’t. He didn’t see this coming because the analysis was too focused on what AI would do, not what AI would make scarce in humans.
The return of craft. AI commoditized middling work, which made truly excellent work disproportionately more valuable. Senior practitioners in almost every cognitive field are doing better than their juniors, and the gap has widened. Harari talked about disruption but undersold the polarization - bottom hollowed, top boomed.
The mid-decade verdict
Harari was right about direction and wrong about magnitude. Reading him in 2026, the broad sweep is useful. The specifics are mostly outdated. The bleakness is overdone.
If you’re a freelancer or solo operator, the practical takeaway isn’t to brace for the apocalypse. It’s to:
- Move up the value chain (more strategy, less production)
- Build trust as a moat - your reputation is harder to replicate than your output
- Get comfortable reinventing your offer every 18-24 months, not once a decade
- Stop reading futurists for tactical advice; read them for orientation, and read operators for tactics
For tactical reading, the things you can only learn by doing is closer to what actually works. Big-picture books are useful in small doses. Doing the work is still the part that matters.
The future Harari described isn’t quite the future we got. The future we got is still strange, still moving, and still mostly favors the people who keep adjusting. The book’s worth reading. Just don’t mistake it for a roadmap.