👋 Hey, it’s Remy.

Quick format update.

A few of you told us the bi-weekly editions were useful, but a bit dense to digest in one go.

We are switching to a weekly rhythm:
one week for a short startup x game theory/maths essay 🧮, the other for community pulse 👾 (wins, asks, behind the scenes, upcoming games).

Same substance. Better pacing.

This week’s piece is about a phase transition that has already reshaped most fields and is about to reshape how startups are built.

At first, people play it loosely. Decisions are driven by taste, intuition, and folklore. Success is explained with stories. Failure is attributed to luck. The field is governed by vibes. It is often defended as something ineffable. “This is an art, you cannot model it. If you tried to formalize it, it would lose what makes it special.”

That belief is usually true at the beginning.

Then something profoundly changes.

The domain becomes legible to maths.
Variables get named. Payoffs get measured. Strategies become comparable.
What used to be fuzzy becomes something you can model, optimize, and compete on.

I call this moment the quantization of a field.

It is a phase transition:
from intuition to structure,
from vibes to variables,
from folklore to systems.

I call the post quantization phase the “GTO era”, borrowing the term from game theory optimal play in games like poker. It is the point where dominant strategies are no longer driven by vibes and folklore, but by structure, models, and equilibrium thinking.

This pattern repeats across history. Here are a few selected examples.

For most of history, navigation was literally described as an art. Seamanship was treated as a craft passed from master to apprentice. Sailors relied on epic stories, memorized star lore, winds, currents, landmarks, and heuristics. Navigation knowledge lived in narrative form. Even classical texts gave navigational guidance as story like instructions such as “keep the Bear on your left” rather than formal procedures.

Early historical accounts explicitly contrast the “seaman’s art” with the later emergence of the “science of navigation”, highlighting how experiential and non systematic the practice once was.

The mathematical foundations of navigation go back much further than the 18th century.

As early as Ancient Greece, geometry and astronomy introduced a formal language for describing position on Earth. Thinkers like Hipparchus and Ptolemy modeled latitude using star angles and treated the Earth as a sphere. Navigation was no longer only folklore. It had a theoretical mathematical layer.

But this mathematics was largely non operational at sea.

The missing variable was time.

Without precise clocks, longitude could not be computed in practice. Sailors could describe the problem geometrically, but could not solve it reliably in real conditions. Navigation remained partly art, partly proto science, but still dominated by heuristics, stories, and embodied skill.

The true phase transition came in the 18th century.

In 1714, the British Parliament passed the Longitude Act, reframing navigation as a solvable, measurable problem.

In the 1750s and 1760s, John Harrison’s marine chronometers made accurate timekeeping possible at sea. For the first time, longitude could be computed reliably. Navigation crossed from mathematical description into mathematical execution.

This is when navigation became legible to optimization.

Once longitude could be calculated, navigation became optimizable. Routes could be planned, travel time estimated, and risk reduced. Seafaring shifted from embodied craft to engineering discipline, enabling global trade and large scale naval logistics.

For much of its early history, trading was framed as a psychological practice. Traders relied on gut feel, narratives, temperament, and “reading the tape”. Market success was attributed to personal flair and an intuitive sense of crowd behavior. The market was treated as a human theatre, not a system.

This intuition driven framing dominated until the early twentieth century. Attempts to model markets mathematically were viewed as abstract and disconnected from real trading.

1900 — Louis Bachelier modeled price movements using probability theory, reframing markets as stochastic processes.

1952 — Harry Markowitz formalized portfolio theory, showing mathematically how diversification and risk could be computed and optimized.

1973 — Black and Scholes introduced formal option pricing, turning derivative valuation into a tractable mathematical problem.

Finance shifted from narrative driven intuition to model driven decision making. Risk, return, and correlation became legible variables. Competitive advantage increasingly moved to those who could design and operate quantitative systems rather than those with only strong instincts.

For most of modern poker history, the game was framed as a human and psychological contest. The edge came from reading faces, spotting tells, projecting confidence, and managing table presence. Strategy culture emphasized intuition, personality, and “playing the man, not the cards.”

Books, televised tournaments, and poker folklore reinforced this view. Great players were described as artists. Success was attributed to feel for the game and mastery of human dynamics rather than formal reasoning about ranges, expected value, or equilibria.

Poker knowledge was transmitted as stories. Hand histories were discussed narratively. Advice sounded like: “he would never bluff there,” “she always folds in that spot,” “trust your read.” The game lived in the space of human interpretation.

Over time, the underlying structure of poker became legible.

Expected value entered mainstream strategy. Players began to think in terms of ranges rather than specific hands. Game theory provided a language to reason about optimal play in adversarial settings.

Then solvers arrived. For the first time, players could approximate equilibrium strategies for complex poker situations. Decisions that were once argued as matters of feel could be compared against model outputs. Lines could be evaluated quantitatively. Leaks could be measured.

Poker stopped being only a story about people.
It became a system with variables, payoffs, and dominant strategies.

At the top level, poker is now played in two regimes:

GTO defines the baseline. It corresponds to a Nash equilibrium of the game: a strategy profile that cannot be exploited by any opponent. Players train with solvers. Strategy is expressed in ranges, frequencies, and expected value. The floor of competence has risen dramatically.

In this regime, there is little room for unconstrained “art.” The structure of the game is largely solved at the level of strategic form. Optimal play is about adherence to equilibrium, not personal expression.

Exploitative play reintroduces degrees of freedom. Pros identify where opponents deviate from GTO and adjust to punish those deviations. What remains of intuition is mostly pattern recognition: mapping behaviors to likely leaks.

This is the GTO era:
not the death of art, but the compression of art inside a legible system.

Across fields, the same transition repeats:

The ordering of steps (2) and (3) is not fixed.

In some domains, the mathematical framework exists before the tooling makes it operational (e.g. navigation: geometry and astronomy existed long before chronometers made them usable at sea).
In others, tooling and data come first, and the mathematical structure is formalized later (e.g. poker: hand histories and solvers came before equilibrium thinking became mainstream).

This is how quantization unfolds across fields.

Startup building has historically been treated as an art.

Taste drives decisions. Things feel malleable. Founders make decisions based on lore: success stories, pattern matching from famous companies, and recycled accelerator “wisdom”.

You hear heuristics like:

“Just build what users want.”
“Follow your instinct.”
“Hustle beats strategy.”
“Do things that do not scale.”
“Make something people love.”

These heuristics (although often contradictory) somehow help.
But they are not models.

They do not tell you why, how much, when, or under what conditions.
They compress complex dynamics into reductive slogans.

What’s changing now is not just that startups are becoming more measurable.
With agents, automation, and software eating execution, more and more of the company is turning into telemetry. Funnels, experiments, unit economics, and workflows are observable in near real time.

But measurement is not understanding.

Dashboards do not give you a theory of the system.
Logs do not tell you which variables actually matter.
A B tests do not explain second order effects or competitive dynamics.

The quantization moment for startups is unfolding right now.
The primitives are appearing. The variables are becoming legible. The structure of the game is starting to surface.

That creates a fault line.

Founders who keep playing purely by feel will compete against founders who model, instrument, and reason about the underlying system. One side is playing stories. The other is playing the game.

This is why I’m unfolding the maths of startups in this newsletter.

Not to replace taste or intuition.
But to make the structure legible.
To turn folklore into variables.
To give language for dynamics that already shape outcomes.

Art does not disappear.
It relocates.

Just like in poker or chess, originality still exists inside a game that is theoretically solvable. But the edge no longer comes from choosing which game you are playing. It comes from seeing the structure of the game earlier than others and learning to play it better.

Startups are soon about to have their quantization moment, and things.

If this lens resonates, I’d love to hear your take.

What part of startups still feels more like art than science to you?

Hit reply and share a thought. We read everything & reply to everyone.

If you want to help shape future essays like this, we’re opening a small number of TechGames Fellowship spots for people who enjoy thinking about startups through the lens of games, maths, and systems.

If this piece made you think, do a friend a favor and send it to one person who should be thinking at this level too. 🧠