Picture someone sitting at a kitchen table after the kids are finally in bed, laptop open, half-drunk mug of herbal tea nearby. For years, she has had a vague idea for a business—custom curriculum design for small learning pods, for example, or a micro-studio creating bespoke art for local nonprofits. She never moved on it. Too many barriers: no time to figure out incorporation, no budget for a web developer, no clue how to do marketing or bookkeeping, no appetite for the legal and tax homework.
But now she types a prompt into an AI assistant.
Within an evening, she has a draft business plan, a shortlist of ideas for company names with available domains, a first version of a logo, a one-page website, basic contract language, a starter bookkeeping system, filled-out forms and instructions for registering her business, and a rough sense of how many clients she’d need to cover her bills. None of it is perfect. But it’s enough to move from daydream to first customer.
That’s the quiet revolution we’re underestimating.
Most of the public conversation about AI and the labor market is fixated on one (very real) side of the story: which jobs disappear, which tasks get automated, which industries will “lose” the most positions.
That conversation isn’t wrong. But it’s incomplete. The same technology that allows big companies to run with far fewer people also lowers the barriers to entry for people who want to create value on their own.
AI is about to pull the labor market in two directions at once: inward, as firms need fewer employees; and outward, as more individuals gain the tools to act like firms.
The coming wave of layoffs
Inside large organizations, the logic is brutally simple. If a machine can do part of a task, fewer humans can do the same job. If a machine can coordinate multiple tasks, fewer humans are needed to manage them. AI turns out to be remarkably good at exactly the kind of work that employed millions of people: following procedures, coordinating handoffs between departments, and navigating bureaucratic complexity.
Some companies will use AI to squeeze costs out of business-as-usual: automating reporting, drafting, customer support, basic analysis, etc. Others will be challenged by newcomers who never built the bulky structures at all. A firm launched in 2026 might not need a marketing department; it has an AI system that writes, tests, and schedules campaigns. It might not need layers of middle management; coordination and monitoring can be handled by software.
Clayton Christensen wrote about “efficiency innovations“—efforts to improve profitability by letting a company do the same work with fewer resources. AI might be the ultimate efficiency innovation. Whether it’s deployed by incumbents to trim fat or by startups that never had the fat to begin with, the destination is similar: less demand for traditional employment inside firms.
We will still have multinational corporations worth billions of dollars. But they will be increasingly lean on staff compared with their 20th-century predecessors: more revenue per employee, more output per headcount, and fewer career ladders.
The personal back office
At the same time, something more hopeful is happening at the edges of the economy.
For most of history, the jump from “I have an idea” to “I have a business” required access to expertise. Lawyers to set up entities and contracts. Accountants to manage books and taxes. Designers and engineers to build products, websites, and marketing. Consultants or mentors to help you avoid rookie mistakes. You either had those skills yourself, had friends who did, or had enough capital to hire them. Many people simply didn’t.
AI breaks that bottleneck. It turns fragments of expertise into something you can “rent by the prompt.”
You still need judgment. You still need creativity. You still need taste, grit, and some tolerance for risk. But you no longer need a small army. The solo founder at the kitchen table has, for the first time in history, a kind of general-purpose back office: a system that can draft, design, summarize, translate, troubleshoot, and simulate at a level that used to require multiple professionals.
Entrepreneurship won’t suddenly become easy. Most new ventures will still fail. Markets will still be unforgiving. Competition may become even more fierce as barriers to entry fall. But the option to try becomes widely available in a way it simply wasn’t before. The barrier shifts from “I can’t even begin” to “Is the potential upside on this idea worth the risk,” which is a very different kind of problem.
The paradox young people will inherit
Put these forces together, and the picture that emerges is neither techno-utopian nor apocalyptic.
Inside firms, AI will quietly erode demand for routine cognitive work. Meanwhile, outside firms, AI will expand the frontier of what individuals can plausibly do on their own or in small teams. That’s the real tension: fewer stable slots in the big machines; more tools to build something of your own.
Whether this becomes a story of flourishing or precarity depends on lots of things—tax policy, social safety nets, and the speed of change. But one piece of the puzzle is squarely in the domain I work in: how we educate young people for the world they’re walking into.
The school of compliance in an entrepreneurial age
For more than a century, mass schooling has been the farm system for large organizations. It has been remarkably good at what it was implicitly designed to do: teach people to be reliable cogs in bureaucratic machines.
The official curriculum covers math, reading, science, history, etc. The unofficial curriculum teaches something else: how to succeed in a rule-bound institution.
You learn that:
- There is always someone above you who sets the assignment.
- The path to success is deciphering what that person wants.
- The safest strategy is to follow instructions faithfully.
- Tasks come with rubrics that specify the criteria for a good performance.
- Your job is to hit those criteria as cleanly as possible.
Do that over thirteen years, and those who get good at winning in the game of school also get very good at reading institutions. They sense where the boundaries are, who has authority, and which boxes need to be checked. They become, in a word, employable—especially in environments where advancement comes from mastering the existing playbook rather than writing a new one.
There is nothing inherently wrong with those skills. For much of the 20th century, this was a rational preparation for a world in which the dominant path to a middle-class life ran through large, hierarchical employers.
But it’s almost the opposite of what today’s entrepreneurship requires.
Innovative entrepreneurship is what happens when there’s no rubric, when no one has written the assignment. When the problem itself is fuzzy, you have to decide which part of it is worth solving. It rewards people who notice friction or unmet needs, test rough solutions, and iterate under uncertainty. It punishes those who are good at execution but expect someone else to tell them what to execute. It favors those who are comfortable with ambiguity and relish innovation. It hobbles those who see their purpose as delivering reliability and efficiency on well-worn rails.
The risk we face is that we will send a generation of students into an AI-transformed economy superbly trained in the old game, just as the old game is shrinking. We’ve taught them to follow procedures, coordinate handoffs, and navigate bureaucracy—precisely the skills AI systems excel at. We’ve led them to expect that career success comes from mastering the rungs on tried-and-true institutionalized career pathways. Meanwhile, the jobs along those conventional pathways are dwindling.
A different kind of preparation
If AI really does reduce the number of people big firms need, while making it dramatically easier for individuals to create value directly, then schools have a choice.
They can double down on being pipelines into a narrowing corporate world—ever more focused on test scores, credentials, and compliance with external standards. Or they can take seriously the task of preparing young people to navigate a world in which many of the best opportunities will be ones they help invent.
That doesn’t mean abandoning core knowledge and skills. Young people will still need to know how to read and communicate with each other and with AI. They’ll still need math and science to conceptually understand how the world works. They’ll still need literature and history to engage with the narratives from the past that define the present. But it also means they’ll need repeated, meaningful practice in:
- Identifying problems that no adult has pre-packaged.
- Spotting unmet Jobs to Be Done where people are cobbling together workarounds.
- Finding their comparative advantages rather than competing on narrow measures.
- Designing and testing solutions that might fail.
- Dealing with ambiguous feedback.
- And exercising agency rather than just obedience.
- Learning how to wrestle with problems that are complex, not just complicated.
Traditional schooling trains students to compete for scarce slots—top class rankings, starting positions on teams, and admission to selective colleges—on standardized dimensions where everyone is measured the same way. That made sense when the goal was landing one of a limited number of corporate jobs. But entrepreneurship works differently. It rewards people who identify niches that are valuable but unattractive to large companies, and who figure out where they can meaningfully differentiate rather than trying to be marginally better than everyone else at the same thing.
My prediction, then, is this:
In the coming years, AI will allow companies to do more with fewer employees. At the same time, it will quietly lower the barriers to entrepreneurship and creative self-employment in ways we are only beginning to see.
The question for education is whether we will keep treating students primarily as future employees of large systems or help them become future innovators in a landscape where powerful new tools of creation are sitting right in front of them.
