On the morning of June 14, Paul Graham posted the transcript of his early-June Oxford Union talk, “How to Earn a Billion Dollars.” Hacker News had it at 478 points and 1,329 comments by evening; as of writing it’s 480+ points, 183 root comments, and 369 unique replies.

Why does a talk on becoming a billionaire produce one of the longest threads in HN history? PG himself wouldn’t be surprised — he opens by saying the talk distills 21 years of watching YC’s 6,500 portfolio companies. But the 1,329 commenters are not buying what he’s selling. So what’s actually under the table?

I read the essay and pulled the three most contested claims, then went through the top long-form replies in the thread. Regular-reader view, no picking a side.

1. PG’s core claim: 93% monthly growth, nine doublings, one billion

The actual content of the essay is a math exercise.

PG opens with a real YC founder: her company grew 93% last month, which means her net worth roughly doubled. He asks the audience to take out their phones and run the doubling themselves.

If you start with $2 million and double it 9 times, you end up with $1 billion.

The math is obviously correct. The thread’s complaint is that “nine consecutive doublings” almost never happens in the wild — each doubling needs another round of capital, another market expansion, another regulatory collision.

Top commenter @naishoya puts it plainly:

He’s saying that the math as he defines it means that YC founders who make less than even millions can anticipate making billions without anticipating doing something unsavory along the way. That’s a fine, and possibly true premise. But, in the real world, as the ’exponential earnings’ stack up, the line between “exponential” and “exploitative” gets blurry fast.

In other words: PG’s math works in a vacuum. In reality, what supports nine consecutive doublings is not just “users told their friends.”

@holistio pushes on the equity dilution angle:

How much do we give to first employees? How realistic is that John Smith, first salesman of the company is getting 2% and should consider himself lucky…

The point: 2% to cofounders, fractional percentages to the first ten, next-to-nothing for early staff — by the ninth doubling, those sub-1% slices own nothing. PG’s “users told their friends” model doesn’t model this layer at all.

2. PG’s counterpunch: “no one cheated” — and the thread brings up Airbnb, Uber, Stripe

PG’s second claim is more aggressive. He calls out a U.S. politician by implication — she said “you can’t earn a billion dollars without doing something bad” — and uses his 30 YC billionaire alumni as counter-evidence.

What she meant was that it’s impossible to get that rich without doing something bad — without cheating in some way.

This is the thread’s biggest flashpoint. @0xbadcafebee just lists names:

Airbnb/Bed Boat, Neighbor, Swimply, Uber/Lyft, Bird/Lime, BlackJet, Waymo/Cruise, Splacer/Peerspace, Zenefits, Tilt, Loomis/Stablecoins, Coinbase, Worldcoin, Stripe, AngelList/Sydecar, Polymarket, Uniswap…

That list is basically the YC alumni directory. PG uses 30 billionaire alumni as “they didn’t cheat” evidence; the thread counters with 30 other YC alumni whose growth stories have a gray side.

@tsimionescu zooms in on PG’s specific example, Airbnb:

So nice of pg to mention AirBnB as one of his examples of what a successful startup who “doesn’t cheat” means. They just were great people with a great idea…

The point: you cite Airbnb as “users loved it” — but Airbnb’s original growth ran on hosts quietly running businesses out of residential properties and skirting local hotel regulations. You don’t get to call that “users liked what we built.”

This single rebuttal turns the whole essay from “math” into “history re-narration.”

3. The most interesting subtext: PG quietly embeds “creative destruction” — and the thread catches him

Late in the essay, PG slips in a line that @AnimalMuppet picks up:

There are other ways to get rich than by starting startups. Some of those do require you to exploit people. But startups are the most common way to become really rich…

This concedes that some wealth paths require exploitation, then cleanly separates “startups” from “exploitation.” The thread thinks the cut is too clean.

@reinitctxoffset writes the longest rebuttal in the thread (2,787 characters):

There’s an older tradition of thought on the matter with better intellectual pedigree and epistemological hygiene: “Behind every great fortune lies a great crime.”

He puts PG’s “exponential math” next to Balzac’s maxim and argues PG’s “users told their friends” narrative deliberately omits the destructive side of creative destruction.

@irishcoffee adds a real-user data point:

Before Uber was a thing, if I tried to call and schedule a taxi pickup in the city I lived in at the time, if they showed up at all they were at least a half hour late. Missed a flight because of it once. I don’t even like uber, but it is objectively a better service.

This comment is interesting because the author concedes the user-side win (faster pickups) but quietly accepts the trade — driver displacement, capital concentration — as a price worth paying. That moral math is exactly the one PG never addresses in the essay.

So is PG wrong?

I don’t think he’s wrong. I think he drew the boundary too cleanly.

His 21 years of YC observation are real: founders do grow from $2M to $1B on “users loved what we built,” and exponential math does compound.

What he leaves out is the other half of the success story: the displaced cab drivers, the squeezed independent hotels, the consolidated publishers, the small developers bought out. The 1,329 comments keep circling back to one thing — PG leaves the “destruction” column of creative destruction for people who are not in the lecture hall.

PG ends the essay with:

“I’d rather be called a SF liberal than a Silicon Valley neoreactionary.”

That’s effectively a preview of how he’ll respond to the thread: he’ll keep his “exponential math + startups are good” frame and own the “Silicon Valley liberal” label. Once that label is on the table, at least a third of the 1,329 comments are arguing about the label itself.

What a regular reader should take from this

This essay is not a startup guide. It is not an anti-PG polemic. It is a debate-opener on what “contribution” actually means.

  • If you’re an indie dev or small operator: PG’s “users told their friends” model is real for you. But watch the boundary — once your growth crosses 100x, the “destructive” side will find you on its own.
  • If you’re an employee or just watching: the strongest rebuttals in the thread are not “PG is immoral.” They are the ones that put “creation” and “destruction” side by side and acknowledge that the user-side win is real. Read @reinitctxoffset, @0xbadcafebee, @naishoya, @tsimionescu, and @AnimalMuppet for that.
  • If you just want to understand why HN blew up: the core is that PG gave the talk in a U.S. election cycle, and the thread will not let anyone stand on a stage saying “we grew by users loving us” while keeping Airbnb and Uber as your poster children.

PG’s essay is worth reading — but please read the thread with it. The five long-form replies above are the actual reading.

References