Earned Delusion

Saudi has billions in venture capital, a genuine need to diversify, and a passionate goal to enable the next generation of Saudis.

Yet, when we look around, we see startups whose ambitions are limited to marginal improvements to regional problems. 

Where are the startups trying to find every vulnerability, make chips an order of magnitude better than Nvidia's, build factories that build themselves, or allow humans to communicate telepathically?

Most of the founders of the ambitious startups I mentioned can't legally rent a car yet.

Imagine your 19-year-old nephew starting a company genuinely believing he can cure ALS or send data centers to space.

Who does that? Who earns the belief, the audacity to tackle such problems?

I think the common theme is: kids who dive really deep into really hard things, and prove time and time again to themselves that they can understand and solve anything.

They spend years in math, physics, and informatics Olympiads. Summers doing algebraic topology, neuroscience research, building flying umbrellas that follow you around. The US produces these kids consistently: the next generation of world-changing professors, researchers, and founders.

When you have the mathematical maturity of a PhD and the confidence — or delusion — of a teenager, frontier work stops being intimidating. You only need to understand the contours of a field before you can start innovating in it.

You don't dare to think of impossible ideas unless you've already spent years proving to yourself that hard problems are solvable, following your curiosity wherever it leads, and building taste for what's important.

That delusion — no, belief — is earned. And it's earned early.

Deedy Das at Menlo Ventures tracked roughly 18,000 IMO, IOI, and IPhO medalists over 25 years. His estimate: they were 4,000× more likely than the average person to found a unicorn.

 The list includes OpenAI, Stripe, Cursor, Databricks, Ethereum, Perplexity, Cognition, Hyperliquid, Scale AI, Decagon, and many more.

Russia on the other hand has schools such as  The Kolmogorov boarding school, Saint Petersburg's Lyceum 239, The D. K. Faddeev Academic Gymnasium who collectively produced Fields Medalists and the only Millennium problem solver to date. Not to mention the Durov brothers, founders of Telegram.

Concentrate unusually capable kids early, allow them to explore ideas and they compound of each other really fast. 

Is Saudi too far behind?

Not at all. Mawhiba has done great work. The fact that students can spend months away from school for Olympiad training shows a willingness to invest in talent rarely seen elsewhere.

The problem is that the parts don't yet compound. Talent is found and trained, then quickly forced down a single rigid, result-based path. Olympiads are powerful because they equip students to explore. Ironically, that is exactly what a Saudi student isn't allowed to do.

The boot camp programs start as early as sixth grade, which is a great thing. The problem is that once you're in, you're expected — actually, forced — to give up every vacation for a camp.

Skip a single camp and, no matter how capable you are, you cannot join again.

The students we equip to explore and innovate are being forced to attend camps they quickly start to resent1

By contrast, my US IOI and IPhO friends spent summers working at startups, diving deep into Russian literature, or even just road-tripping along the West Coast.

The point is that everyone who eventually goes to camp does so willingly, passionately.

Po-Shen Loh coached the US IMO team through four first-place finishes, yet he was pretty explicit that he did not care about medal results. He cared about who the students became 20 or 30 years later; the point was to make them genuinely curious, not better at collecting medals.

The effort we're putting in feels like pushing a car that's still in park.

We need more paths, ones where the goal isn't competition results but to enable those with the most potential and drive to explore.

Doing abstract algebra, quantum physics, or mechanistic interoperability as a kid gives you the confidence and ability to create the next Periodic Labs or DeepMind. 

It's not just theoretical ideas, a kid who spends years making real things collects physical proof that hard ideas can become real.

The same thing happens outside Olympiads. Some of my friends at MIT spent their teenage years building:

  • Liong Ma started tinkering at 12: a pressure-sensor robotic arm, a cheap seismometer and spectrometer, then an electric race car with over 40 circuit boards, 20 controllers, and roughly 6,000 lines of code.
  • Vaughn Khouri welded a 45 mph go-kart, built an electromagnetic Captain America shield, then made a power-generating seesaw and a suction-mounted workout machine.
  • Anhad Sawhney’s portfolio includes a six-week solar-plane project, open-source micro-drone swarms, and a nuclear fusor.

You can teach yourself physics, design the board, and weld the frame if you just persevere enough. 

Of course such a change doesn't happen overnight, but I propose a few concrete steps.

  1. Fund passionate professors and researchers to run programs around whatever they'd love to share to curious kids, teenagers, and adults alike

    The Nesin Mathematics Village offers lecturers room and board, but does not pay them; they travel there themselves.

    • Max Dickmann — Paris Cité and the Sorbonne — spectral spaces.
    • Roman Kossak — CUNY Graduate Center — descriptive set theory.
    • Sten Kaijser — Uppsala — orthogonal polynomials.
    • Alp Müyesser — Oxford — extremal combinatorics.
    • Sven Gnutzmann — Nottingham — quantum mechanics.

    Give serious professors a room full of genuinely curious students and many will come. The chance to teach what they love to kids who actually want to be there can be enough.

    And there is no reason to limit this to Saudi students. Saudi should sponsor these rooms wherever they can exist, putting its students inside a global network of unusually curious peers and professors.

  2. Sponsor far more seats at programs like Ross, PROMYS, Mathcamp, and RSI. We already sponsor two or three RSI seats a year, surely we can afford to do so for similar programs too 
  3. Make Olympiad camps optional — and therefore intentional — and rerun selection every year. Missing one summer should never become a permanent exit.
  4. Microgrants of $5–10k to let kids build something real, paired with equipped maker spaces in every major city. No pitch decks, no demo days; just enough money and equipment to try.
  5. A 10–15 student fellowship: one serious mentor per student, funding to travel and build, and no predetermined track.

Saudi already finds the kids and spends the money. The missing piece isn't another camp. It's giving them enough freedom, serious peers, and hard things worth doing for the belief to compound.

1 This was not a formal survey: I asked more than 30 students, across several Olympiad camps, how the training had changed how they felt about their subject. All but one said they had lost passion for it; some said they had started to resent it entirely.