Sal Khan on "Job Shock": Why the Real AI Opportunity Is in Boring Industries

Sal Khan has spent his career being an optimist. The founder and CEO of Khan Academy, whose platform teaches 100 million people a year and which built one of the world's first AI tutors, is not someone prone to doom. So when he says his next book is called Job Shock and that he's "increasingly worried," it's worth paying attention. In a conversation with Silicon Valley Girl, Khan lays out what AI is about to do to work — and where the genuine opportunities lie for anyone willing to look past the crowd.

This post condenses the episode. The claims, numbers, and predictions are Khan's.

The Wake-Up Call

Khan's shift from optimism to concern has a specific origin. A friend from college, now a prominent VC, told him that one of his portfolio companies had automated 80% of a Philippines call center — hundreds, maybe thousands of jobs — with a single generative-AI solution. Khan points out that business process outsourcing, mostly call centers, accounts for 5 to 7% of the Philippine economy. If this becomes a trend over the next three to five years, he asks, what will all those people do?

His broader argument, first made in a New York Times op-ed, is that we can't just shrug and say "the industrial revolution turned out okay." That era, he notes, also produced the bloodiest century in history, world wars, and disastrous experiments with communism. AI's disruption, in his view, demands an actual plan — which is what Job Shock is about.

Who's at Risk, and Who's Safe

Khan expects the dislocation to become real within five to ten years — slower than the AI evangelists predict, but fast enough to arrive "before your kids go to college." Driving worries him most: it's one of the largest employers of men globally, and mass male unemployment, he warns, breeds anger, polarization, and a drift toward political extremes.

Beyond drivers and call-center staff, he says "the writing is on the wall" for software engineering, design, and product management. These roles aren't necessarily vanishing, but they're blending together and changing fundamentally — even at Khan Academy, where designers and product managers are now being given full development environments to ship code. The people genuinely at risk, he stresses, are those who don't want to adapt.

The safe jobs, by contrast, lean into the human element. Khan names teaching — not the "sage on the stage" dispensing information, but the coach, architect, and motivator who runs a Socratic dialogue and makes eye contact with students. He adds nursing, hospitality, and relationship-based sales, where the best salespeople are trusted advisers who'll tell you their product isn't right for you. Anything that can be done purely on a computer, with no human connection, is under threat.

Why AI Won't Trigger His Layoffs

Khan Academy has 350 employees, two-thirds of them in product, and Khan admits they're nervous: if AI makes them three times more productive, won't two-thirds be cut? His answer is emphatic. "If we could do three times more with the same resources, we will do three times more. If we could do 30, we'll do 30." Layoffs, he says, come from a shortfall in philanthropy or revenue — never from AI, which only lets the organization do more. But he's been equally clear that jobs are evolving, and Khan Academy has added "learning new tools and adapting" directly into its career rubric.

Internally, the change is dramatic. Engineers now run five to ten coding agents at once, and Anthropic reportedly ranks Khan Academy among its heaviest users of code-review agents. The bill is eye-watering — about $1.2 million a year for roughly 200 engineers, with one engineer spending $3,000 of compute in a single day to accomplish what would normally take three or four months. Khan's take: at $1–3 million in tokens to produce what used to cost $20–30 million, "we should keep doing it." He estimates the organization now moves 50 to 100% faster, and it's accelerating. A feature once slated for "next school year" was vibe-coded at a hackathon in a day and shipped in a month.

He's also using AI as a kind of chief of staff. With connectors to Slack, Gmail, and docs, he asks it daily "what's falling through the cracks?" and lets it draft — never send — replies and proposals. Something that once took a week or two now has a first draft within ten minutes of a meeting. His discipline: keep a human in the loop, and always verify sources (an IRS.gov citation, say) before publishing.

Can AI Just Build Another Khan Academy?

Asked whether AI might simply replace his organization, Khan gives two answers. The honest, ego-driven one: yes, he worries, and in three to five years it may be plausible. The "correct" one, which he tells his team constantly: as a nonprofit, if someone builds something as good or better and proves it works, they should celebrate — one of the world's problems just got solved. And in any case, the real moats aren't easy to vibe-code in a garage: efficacy data, rigorous psychometrics, and the trust and data-privacy relationships with school systems all take years to build.

Reinventing the Degree

Khan's constructive response to Job Shock is the Content Institute, launched with TED and ETS. It's an accredited bachelor's and master's degree for a maximum of $10,000 — he intends it to cost considerably less. The structure has three parts: real academics (history, civics, accounting, statistics), online group simulations where students do genuine work (build a business in four hours: prototype, ad, business plan, customer surveys), and peer review of how they performed. Within two to three weeks, 3,000 people had signed up — most already holding bachelor's or master's degrees, driven by a palpable fear that they need to requalify.

The payoff is a different kind of résumé. Students complete 20 to 50 rigorously designed, rated project simulations — co-designed with McKinsey, Bain, Google, Microsoft, Replit, and Accenture — that surface what Khan calls the "durable skills," ETS's five C's: communication, collaboration, creativity, critical thinking, and leadership. Instead of a GPA that merely signals "probably smart," employers get evidence, including video artifacts, of how a person actually communicates and collaborates. Khan's goal is that a top-10%-in-communication Content Institute graduate gets treated as seriously as a top-10% graduate of an elite university — and that the stigma around online education disappears.

Is College Still Worth It?

Khan is refreshingly candid here. Elite college, he says, is now a "luxury experience" — one he'll happily pay around half a million dollars for his own son, a high-school junior, to have, for the friendships and coming-of-age it offers. But he's clear-eyed that a $10k competency-based degree could take two years instead of four, freeing up two more years to earn. And he floats genuinely different uses of that money: send your kid to travel the world reading great books; seed a business; put it toward a house; or even buy a $200,000–$300,000 unautomated mom-and-pop business and have your kids automate it, private-equity style.

The Real Opportunity

Asked what he'd do if he were a 22-year-old MIT grad today, Khan delivers the line that gives the episode its title. "Don't run to where everyone else is running. Try to find the lanes that are most empty — the most boring industries that are most ripe for applying these technologies." For any individual worried about being replaced, his prescription is practical: get over the "activation energy" of trying new tools, look at your own workflows, and automate them incrementally — "none of this stuff is rocket science."

And differentiation, he insists, is still easy. In 1996, saying "I have a web page" got you hired. Today, the equivalent is showing up and saying "I come with 100 agents," or "here's a Loom of a day in my life where I've automated my work with agents." Someone like that, Khan says, he'd hire at Khan Academy tomorrow. His final caution is a telling one: AI is a flatterer that will call you a genius, so he's learned to tell it, "be critical of me" — and even then, it praises him for wanting the pushback.

---

Originally published on Silicon Valley Girl. Watch the full episode: https://www.youtube.com/watch?v=s-Iz-lLAhdg