Startup School

Patrick Collison on Why It's Never Been a Better Time to Start

Patrick Collison· CEO at Stripe
·~31 min·English·Y Combinator
Business StrategyAgentsAI Company
TL;DR

Stripe's CEO argues the AI era has lowered the cost of starting a company and widened who can win — making the classic fears, and the classic lean-startup playbook, look overstated.

01Core Mental Model

Cognitive L1 Cache

In the AI era, the knowledge you hold in your head is your L1 cache — reachable far faster than anything you have to stop and ask a model for.

That's a hell of a lot slower than knowing it in cognitive L1 cache.

— Patrick Collison, Startup School
Key Insight
The revealed-preference tell is that the labs themselves still pay a steep premium for raw cognitive ability — so outsourcing your own learning before that premium disappears is a bet against the very companies building the models.
02Worldview

Take the Under on the Underclass

The fear that missing this moment traps you in a permanent underclass is the same millenarian hype every big leap attracts — and Collison takes the under.

I would take the under on this being the last couple of years to get a company going.

— Patrick Collison, Startup School
Key Insight
Collison isn't dismissing AI's impact; he's separating the technology's real weight from the social panic around it — the aviation age was genuinely transformative and still under-delivered on its loudest prophets.
03Origin Story

Squirrels in a Trench Coat

Stripe looked absurd — two kids doing financial services before 'fintech' was even a word — and survived because it was built on a concrete problem people would pay to fix.

the fact that it was grounded in such a concrete actual real user problem saved us.

— Patrick Collison, Startup School
Key Insight
The lesson isn't that audacious ideas win — it's that a viscerally-felt, pay-to-fix problem is the ballast that keeps an implausible company from tipping over when everyone around you is looking for the button to call security.
04Key Decision

Larval, but Live

Stripe waited two years to launch publicly, but had live production users from month two — so it was learning from reality the entire time.

every week, we had actual customer feedback, requests, new users coming in. We're learning things from reality as opposed to our own kind of hypothesized or extrapolated conception of it.

— Patrick Collison, Startup School
Key Insight
'Launch early' versus 'launch late' is the wrong axis — the real variable is contact with real users, and Stripe had that from month two, which is why a two-year build was never a two-year guess.
05Startup Strategy

Decorrelate in the Era of AI

The classic lean-startup playbook still works, but the era of AI rewards starting broad and ambitious instead of hiding in one narrow niche.

now I think you can start these much more aggressive and ambitious things up front.

— Patrick Collison, Startup School
Key Insight
If everyone runs the same lean playbook, the niches get tilled flat — so the edge moves to founders willing to decorrelate, and cheap AI-built software is exactly what makes an ambitious version one affordable.
06Competitive Fear

The Labs Aren't Omnipotent

The fear that a lab will trample your idea is overstated — no organization executes a hundred priorities well — but improving models will still obviate specific tasks.

Will rapidly improving AI capabilities do this or will the labs specifically themselves do this?

— Patrick Collison, Startup School
Key Insight
Collison splits a single anxiety into an org problem and a capability problem — the org problem shrinks with scale as coordination cost rises, while the capability problem is the one actually worth forecasting.
07The Data

Never a Better Time

By every metric Stripe can see — new businesses up roughly 2x year-over-year, the median doing better, time-to-revenue falling — it has never been a better time to start.

by all the kind of objective metrics we can look at, uh, it seems to be a better time than ever to start a business.

— Patrick Collison, Startup School
Key Insight
Collison carefully hedges the future, but the convergence of independent signals — median performance up, revenue-threshold odds up, and time-to-revenue down — is far harder to dismiss than any single headline number.
08The Big Picture

Thousands of Winners, Not One

Stripe's data points away from a few firms swallowing the economy and toward many thousands of winners — a more decentralized, broad-based prosperity.

I think there are going to be many thousands of winners.

— Patrick Collison, Startup School
Key Insight
A payments company sees the whole cross-section of the economy at once, so 'thousands of winners' isn't optimism — it's what the transaction data looks like before any narrative gets imposed on it.