The Death of the First Job
Why the next generation may build a business before it builds a career, and what it will take to make that a real pathway rather than being a lucky exception
Opinions expressed by Entrepreneur contributors are their own.
You're reading Entrepreneur Middle East, an international franchise of Entrepreneur Media.

In March 2026, MyFitnessPal, a calorie-tracking app with more than a decade’s head start, bought a rival run by two teenagers.
Zach Yadegari and Henry Langmack built Cal AI while still at Roslyn High School on Long Island. The app does one thing. You photograph your plate and it estimates what you are about to eat. They did not invent the AI behind it. They packaged it, put a subscription on it, and sold it through app stores that reach almost anyone with a smartphone.
By the time of the sale, MyFitnessPal said, Cal AI had passed 15 million downloads and more than US$30 million in annual revenue, in under two years. The company was seven employees and a few contractors. MyFitnessPal kept every one of them.
By October, Yadegari, now 19, had left the acquirer and raised US$10 million for his next company.
It would be easy to file this under prodigies. Yadegari was coding as a child and sold his first website for US$100,000 at 16. Most 19-year-olds will not do what he did, and no sensible policy is built around outliers. Closer to home, we can find similar attempts being made by the youngsters. Students and young residents in the UAE have been starting ventures of their own, from GoPool, the school carpool app that preceded Zepto, to newer efforts such as Mengo Engine, Carbon2Capital and Reach LLM. Most will not become Cal AI and that is not the point. The point is that a growing number of young people here are treating a venture, not a job application, as their first move.
What interests me is less the talent than what the story leaves out. There was no office, no sales team and, crucially, no first employer. Nobody hired Yadegari to learn the trade; he learned it by selling to strangers. That part is no longer unusual, and it is what this article is about.
Two doors, moving in opposite directions
Look at three very different labor markets and the same pattern appears: the way into a first job is narrowing, while the way into a first enterprise is wider than it has been in a generation.
| UAE | Europe | United States | |
| Young people out of work | 5.2% of 15–24s(2024 LabourForce Survey) | 15.5% of under-25s, EU(Sept 2026) | 5.7% of recent graduates aged 22–27 (Q1 2026), against 4.2% for all workers |
| Early-stage | 19.2% of adults | Germany 13.0%, its highest since 1999 | 17.7% of adults |
| Entry-level signal | ~250,000 new companies formed in 2025 | UK graduate vacancies down 45.6% year on year (July 2026) | 22–25s in AI-exposed jobs 19% below peers in less-exposed jobs (Aug 2026) |
The measures are not like-for-like, since each country counts youth, graduates and founders in its own way. Still, reading across the rows, three things stand out.
In the United States, recent graduates are now more likely to be unemployed than the workforce as a whole. For most of the past three decades it was the other way round. Roughly 41% of those who do work are in jobs that do not need their degree.
In Europe, about one in six under-25s is out of work. Yet Germany, hardly a nation of risk-takers, has just recorded its highest startup rate in more than 25 years of measurement.
The UAE’s problem is a different one, and in some ways more interesting. It is not sitting on a pool of idle graduates. Its challenge is what happens to the businesses it is so good at creating. The latest Global Entrepreneurship Monitor found startup activity at historic highs worldwide, while progression to established businesses is stalling in many economies. The report calls this a survival gap.
What the first job was really for
We talk about the first job as if it were mainly a salary, but it did four things at once, and we are only noticing them now that they are coming apart.
It paid you, which bought time to get better. It taught you, through supervisors, repetition and the slow accumulation of small mistakes that someone else absorbed. It connected you, to colleagues, clients and the people who would later hire you again. And it certified you: a recognized employer on a CV told the next employer you could be trusted with real work.
The Stanford research on early-career workers points at exactly where this is breaking. Entry-level hiring has weakened most in occupations built on codified knowledge: the documented, teachable tasks that once made up a junior’s first two years. That is precisely the work AI now does quickly and cheaply. The tasks are disappearing, and the teaching that was bundled with them is going too.
Which leaves a harder question than whether young people will find jobs: if the first job no longer supplies those four functions, what will? Any alternative worth taking seriously has to answer that.
The First Enterprise Economy: A working framework
I use the term First Enterprise Economy for an economy in which a young person’s first sustained paid work can credibly be their own enterprise, and in which that enterprise does the four jobs the first job used to do.
The definition is deliberately demanding. A country does not qualify by issuing more licenses, or by counting gig work as entrepreneurship. It qualifies when a 22-year-old’s first enterprise pays, teaches, connects and certifies about as reliably as a graduate scheme. I find it useful to break the idea into two pieces, each of which can be used to test any program that claims to support young founders.
Does it pay, teach, connect and certify?
Every first-enterprise program can be judged by asking how it replaces each function of the first job, and where it is most likely to fail.
| Function | How the first job supplied it | How a first enterprise must supply it | Where it usually breaks |
| Pay | A monthly salary | A first paying customer, then repeat revenue | The months before the first invoice |
| Teach | Supervisors andrepetition | Customers, mentors andcohorts of peers | Mistakes are expensive andnobody explains them |
| Connect | Colleagues and theemployer’s clients | Buyers, platforms and peer networks | Isolation; the founder works alone |
| Certify | A recognized employer on the CV | A verifiable record of revenue, clients and delivery | Employers and banks don’t know how to read it |
The last row is the most neglected. A young founder who has served 40 clients in three countries has done more than most graduate trainees. Yet hardly any recruiter, bank or visa office has a standard way to recognize it.
What a first enterprise stands on
Underneath any first enterprise sit five layers of support, and most policy stops at the first one.
1. Permission. Legal status: a license, a permit, a bank account, a tax number. This is where governments have done the most.
2. Tools. AI, cloud software, payments and digital storefronts. These are now cheap and largely global, which is what makes the whole idea possible.
3. Market. A first customer. This is the hardest layer and the least supported. Capability does not create demand.
4. Money. Working capital sized to the business: small, fast, and repaid from revenue rather than secured on assets the founder does not have.
5. Judgment. Pricing, contracts, compliance, when to say no. The craft knowledge a good manager once passed on without anyone noticing.
A useful exercise for any ministry, university or bank is to ask which layer it actually owns, and who owns the other four.
Where the UAE stands
Measured against the five layers, the UAE is far ahead on Permission, and Tools — AI, cloud software, digital payments — are now available to anyone.
On Permission, the record is hard to argue with. Around 250,000 companies were set up in 2025, taking the total past 1.4 million, and the government is aiming for two million within a decade. In Abu Dhabi, freelance licenses rose 261% year on year in the first quarter of 2026. For the fifth year running, the Global Entrepreneurship Monitor ranked the UAE first of 53 economies for the quality of its entrepreneurial environment. It is also one of the few places where new founders expect AI to matter a great deal to their business.
Having led the world on Permission, and with Tools now available to everyone, the UAE is unusually well placed to lead on the next three layers as well: helping young founders find their first customers, access capital suited to their size, and learn from people who have built businesses before them. Few countries are in a position to attempt this. The UAE is one of them. Some of those activities are ongoing but would need more integrated effort to make them a success.
The objections I take seriously
The obvious objection is survivorship bias, and it is fair. For every Cal AI there are thousands of apps nobody downloaded, which is why success should be judged by survival and by what happens to founders afterwards, rather than by launches. The goal is not more prodigies but a higher floor for the ordinary first enterprise.
Nor will this work for every profession. Nobody should open a solo surgical practice at 23, and much of law and structural engineering needs years of supervision. The First Enterprise Economy is a second way in, alongside the first, and it matters most where the work is already digital and the customer can be anywhere.
The evidence on AI is also genuinely contested. A September 2026 NBER paper found no significant AI linked rise in graduate unemployment, and the Stanford team now dates the clearest effect to 2024 rather than 2022. My argument does not rest on AI being the main cause, though. It rests on the first rung narrowing, which the table above shows on almost every measure.
The criticism that bothers me most is that all this is precarity with a better name. A freelance permit can hide underemployment as easily as it can launch a career. The only real protection is honest measurement of revenue, of survival and of what the founder does next.
The milestone we should be counting
Go back to Cal AI for a moment. Nobody gave those seven people an offer letter; their first professional milestone was an invoice, from the first stranger willing to pay for something they had made.
Most young founders will never get near US$30 million, and they don’t need to. What they need is what Cal AI had, at a smaller scale: a customer, a way to get paid, someone to learn from, and a record that the next employer or lender can read. Most nations should endeavor to provide them to its youngsters.
Countries that build those four things will not be judged by how many entry-level jobs they protect. They will be judged by how many of their 23-year-olds can say, truthfully, that someone is already paying them for their work.
The offer letter is not dead. It just isn’t the only way in any more.

In March 2026, MyFitnessPal, a calorie-tracking app with more than a decade’s head start, bought a rival run by two teenagers.
Zach Yadegari and Henry Langmack built Cal AI while still at Roslyn High School on Long Island. The app does one thing. You photograph your plate and it estimates what you are about to eat. They did not invent the AI behind it. They packaged it, put a subscription on it, and sold it through app stores that reach almost anyone with a smartphone.
By the time of the sale, MyFitnessPal said, Cal AI had passed 15 million downloads and more than US$30 million in annual revenue, in under two years. The company was seven employees and a few contractors. MyFitnessPal kept every one of them.