All jobs are exposed to natural selection; at work, if you do not evolve, you are doomed to be replaced by the intern, by the market, by new technology, or all three. But some jobs are far more exposed to creative disruption than others, and the champions of these professions exhibit a pattern worth studying.

House Odds

Take academia; around 4% of the population of the UK and US embark on a PhD, yet 50% fail to receive a doctorate, 70% won’t stay in academia for five years, 97% do not achieve tenure or permanent positions, and 99.6% won’t become professors. One in two hundred of those who start manage to persist—let alone have a signficant impact on the profession.

Take sports: every European seven-year-old is forced onto the football pitch, yet only 572 persist to the premier league. The same principle applies to music, comedy, acting and software engineering. For those who try and fail to do something great, falling short is rationalised as some deficiency of either hard work or “natural talent”.

But the most overlooked property of any system that persists is the heuristic instinct that is carved out during competition. This is what most distinguishes the master from the amateur. These instincts let your supervisor effortlessly poke holes in the paper you sunk hundreds of brain-cycles into, or why your team breaks under pressure against a better side.

Maximising Heuristic Instinct

Heuristic instinct means that whover has the most accurate predictive model wins.

Take the brain of a savant for instance: able to run advanced pattern-matching without explicit training, like Yashantha from this famous story. These characters inherit default wiring to closely approximate the type of pattern recognition needed to perform a specific task well. For competitors, this superpower feels profound as a teen, but in the long-run, anyone with a plastic brain can learn these tricks more acutely.

The savant’s savant operates across many domains: the polyglots who absorb languages with perfect fidelity, the poker champions who dominate every game at the casino, or the Renaissance men of art, engineering, and science. These are the fittest members of the population with the greatest means to outgrow all competition in the long-run. [1]

Yet, even these figures are forced to specialise. With just one brain, one lifetime, and a family to tend to, your ability to go deep to beat your competition deteriorates as you accrue more draws on your attention. Even the best can only hope accrue PhD-level insight in one to three fields, without trading desirable human attributes like a loving home life, or close friends.

For the modern day renaissance man, this means your best shot is to specialise in the field that most intensely develops your general heuristic instinct. Traditional paths—researcher, architect, surgeon, trial lawyer, quant—provide security and prestige, but they constrain the feedback loops that sharpen your instincts; you’re sheltered from the market, you’re sheltered from the technology, you are required by your employer to fit to a specific domain at the expense of more general heuristics.

This limits your heuristic development in the long-run. When your supervisor can afford to be patient, he creates an abstraction over reality that feels like a feature, when it’s more like a bug. The boss who gives you nine strikes is ultimately depriving you of the more direct paths to mastery.

But fear not, for there exists an opposite extreme: a world where you receive no buggy abstraction layer, no buffer against reality—a world where the feedback is so honest, brutal and immediate that only the antifragile persist.

Extreme Entrepreneurship

Technological entrepreneurship represents the most intense evolutionary cycle of all human endeavour. It is directly involved in the world of markets, the world of science, and the world of social status. In software, often, the half-life of any competitive advantage approaches zero in the order of days. The steeper this gradient, the more forcefully your heuristic instincts will develop.

This compounds when your competition consists of multi-brained intellects; massively parallelised organisations moving with bandwidth in the order of hundreds or thousands the means of any given founding team. This makes the myth of the startup founder aggrandising; at the helm of a technology juggernaut you possess the mandate to steer humanity and carve the world to your will.

When your existence hinges directly on learning what is wrong before it bankrupts you, those who persist become antifragile. This is why successful business leaders so gleefully describe running companies with in violent, vicious terms. Where else is winning described as “like eating glass; you just start to like the taste of your own blood”?

The heuristic development of an entrepreneur of 2 months, 2 years, and 20 years is staggering. As a consequence, second time founders raise dramatically more money than first-time founders because (1) there aren’t many of them, and (2) those who succeed have strong heuristic instinct.

As each new generation of Zuckerbergees compress the build cycle further, I wanted to document the intractable path from conception to success. Even the best entrepreneurs must climb the five levels of heuristic instinct:

Level I: Passion-Builder

The Tinkering Entrepreneur

This is where most startups are born, and where most startups die. We will characterise this level as passion building if it originates from a genuine curiosity about hard, unsolved problems. Many good startup stories are described as successful tinkering: technical students discover something that they are uniquely capable of solving, write a search engine, and discover intense consumer demand.

So what seperates the should-be-avoided tinkering from the productive?

Bad passion-building is distinctive: the Notion founders spoke about how their first product was an indulgent, feature-rich undertaking built for themselves that didn’t fair well with others. Only when they rebuilt the project around a smaller, more obvious wedge were they able to find traction and persist long enough to build a super-app people wanted.

At its worst, passion-building is procrastination disguised as innovation. It is the delusion that anything except user traction creates businesses, it is the myth that in order to compete with a billion-dollar juggernaut you must create something worth a billion dollars out of the box, or the belief that nine months spinning code without anyone breaking your door will result in great traction once you’re finished. The reality is, being technical is not enough to build a great product.

So what about those who succeed?

The HelixDB founders built a graph-database from scratch because other options were terrible, tickled the interest of a small community of enthusiasts and sold the vision to heavy-pocketed enterprise. The early Stripe team started building payment infrastructure because existing solutions were inadequate. They tested this with a few technical users, who soon became early employees building a product they would pay for more. Both engaged a small following and validated usage with a small number of committed users before scaling. The same story goes for most successful companies: Airbnb, Slack, TaskRabbit, Loopt. These builders pushed into uncharted territory with hypotheses that were formed, tested, and falsified on repeat. They built something with 1% idea and 99% iteration: starting with a hunch and following the signal thereafter.

Geoff Hinton once said that the best behaviour a student can exhibit is extreme skepticism about every new piece of information they learn; this way, they will trust only what they can prove, and in the long-run, earn a very strong heuristic indeed. The same is true for entrepreneurship, where self-belief borders on delusion until people pay you to keep going. While the most successful passion-builders are testing hypotheses with code, doomed passion-builders are nothing more than headless visionaries.

The strongest advice I have found is to actively constrain myself from throwing the kitchen sink into a product. Build something simple that drives at the essence of the problem you are solving, and then commoditise your complement. For software, that means one screen with one core action until people are practically begging you to build more.

Level II: Hustler

The Forceful Entrepreneur

At some point, you launch and discover that users don’t care about you. From here, entrepreneurships starts to feel real. Manually onboarding and keeping new users occupies this space, and for shy founders who write code, this does not come naturally.

Sales and persuasion come most naturally when you are selling something you use every day, but it helps to develop some social growth hacking techniques just to be safe. “Doing things that don’t scale” is the mantra of founders in this phase: start with anybody, as long as it isn’t yourself (or your “Mom”, in most circumstances).

Showing your peers something you built is the most natural way of hustling. Demoing your build on X is a surefire way to see if anyone is interested. The Stripe team would acquire users via “Collision installations” where they would take the laptop of users and set them up on the spot. Airbnb flew around the country each week going door to door, recruiting new users and helping existing ones improve their listings.

Every passion-builder that went on to run a successful business successfully made this transition. It is the transition towards product market fit, towards Ramen profitable, and it is characterised by relentless iteration.

Level III: Cook

The Momentum Entrepreneur

When inbound consistently grows faster than outbound, you have made it to the 1% Club of entrepreneurship. This is where entrepreneurship takes on its more characteristic dimension from the outside. If you can crank the hog hard enough, it will being running on its own.

Managing your inbound efficiently becomes more essential once you make it to this level. This isn’t trivial; only the best in each YC batch will fall under this category for any meaningful duration. Shotgun launches might give you a waitlist and a surge of press coverage, but this will dry up unless you can continue to be great.

Entrepreneurs in this domain have discovered a product that consistently solves a problem people actually pay for. Maintain this for long enough and soon you’ll be able to employ people without taking on debt, or announce a more grueling release schedule with greater up-front investment. In short, you’re gaining control over your domain of influence now that your Ideal Customer Profile is understood more deeply, carved out over genuinely rigorous testing.

At last, metrics dashboards like PostHog track heuristics more explicitly than vague social heuristics (although nothing will ever completely substitute vibes). Entrepreneurship starts to feel more scientific than opportunistic.

Level IV: Thought-Leader

The Master Entrepreneur

The most competitive Level III entrepreneurs become masters of their subdomain. With sustained momentum, your company has evolved into a later stage startup or established tech company.

Perhaps close employees feel like they have a PhD supervisor when working with these characters when they “see around corners” and know what is worth testing and what isn’t.

An entrepreneur at this level is not just finding product-market fit in existing markets, they’re creating new markets happen or reshaping existing ones. “What valuable company is nobody building?” becomes central to the thought process.

Most of the YC Partners appear to be engaged in the behaviours of Level IV founders. This is the behaviour most aspire to attain when starting out in business, knowingly or not. You have enough customer signal and market understanding to see opportunities others miss.

The best founders operate here sometimes for years, and few remain at this level for long. Some entrepreneurs that operate at this level include:

  • Marc Benioff - Salesforce (1999-2010) - Created the “Software as a Service” category, evangelised cloud computing
  • Tobias Lütke - Shopify (2015-now) - Created “democratized e-commerce” category, enabled millions of merchants
  • Daniel Ek - Spotify (2008-now) - Changed music consumption forever, created streaming economy for artists
  • Brian Chesky - Airbnb (2014-now) - Trial-and-erroring beyond core product into “Experiences” and other travel categories
  • Travis Kalanick - Uber (2009-2017) - Created the ride-sharing category, expanded globally despite regulatory battles (ended due to governance issues)
  • Patrick Collison - Stripe (2010-now) - Simplified online payments, became developer-first payment infrastructure
  • Nat Friedman - Xamarin (2011-2016), GitHub (2018-2021) - Aggressive and visionary product release schedule at Github, scaling to 73M+ developers as CEO

Level V: Aura Farmer

The Meta Entrepreneur

You’re doing things no one else could do sheerly because of your reputation and visionary powers. You have a track record of building great products, perhaps across multiple companies, and still you keep building and pushing the boundaries further.

The main differentiator between you and the Level IV cohort is the degree to which your “higher level vision” for humanity captures the mind of the public. If your tweets could shape whole cultural narratives, you’re firmly within this category.

This is most obvious when you have a personality people immediately recognise, when you actively manufacture memes and frameworks that Level IV founders and the broader community adopt.

You are in the business of shaping reality to your will. You could drop in on almost anyone and it would be received with a mixture of awe and intrigue. Level V founders include:

  • Steve Jobs (1998-2011) - Apple’s design philosophy changed entire industries beyond technology
  • Elon Musk (2006-present) - Tesla, SpaceX, Neuralink, xAI - Can announce projects and instantly attract talent/funding others couldn’t access
  • Sam Altman (2019-present) - OpenAI’s AGI mission, can raise billions based on vision alone
  • Larry Page & Sergey Brin (2004-2015) - Google’s “organize the world’s information” mission shaped the internet era
  • Mark Zuckerberg (2021-present) - Meta’s metaverse bet and open-source AI strategy with Llama
  • Jeff Bezos (2000-present) - Amazon’s “everything store” strategy, Blue Origin - Reputation allows entry into any market

Bonus points if you founded a trillion dollar company.

Double bonus points if keep buying companies and shaping them like a founder would.

Understanding this hierarchy helps explain how certain founders are likely to behave depending on where they are in the cycle.

Notes

[1] This quote from Paul Graham comes to mind: “Sam Altman has it. You could parachute him into an island full of cannibals and come back in 5 years and he’d be the king.”