It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest.
Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations, 1776
Before PrivaPaid existed, I was experimenting.
I had a payment rail, Lightning invoices, settled without custody, and I wanted to see what happened when I dropped it into a real e-commerce flow on someone else’s site. The simplest way I could think of was an iframe. The iframe shows a Lightning invoice; the customer pays; the iframe tells the host site the payment cleared.
I could have done this the way every other payment integration does it. Push a customer object into the merchant’s account system, name, email, billing address against an order ID, the way every checkout the merchant had ever shipped expected. I started looking at how to do that and I noticed, before I had built any of it, that I did not have to. The iframe already had everything. The Lightning invoice was the credential. The payment was the receipt. The amount was the price. There was no missing field on our side. The customer object the merchant’s stack expected was something the merchant’s stack expected because the stack had been built around the assumption that the field had to exist.
So I let the iframe collect what it actually needed, and I asked for customer info only where delivery required it. A physical good gets a name and an address. A digital good does not. And once I sat with what that meant, that for any digital good small enough to be casual, a buyer would happily skip identity rather than hand it over for the price of an article, the use cases lined up faster than I could write them down. A photographer selling one print to a stranger. A reporter in another country selling an article to a reader who would never have a payment relationship with her newspaper. A musician selling a track to fans across borders that block the usual rails. Small risk on the buyer’s side, no relationship debt on the seller’s, no platform in the middle holding either of them by the wrist.
I was not solving an integration problem. I was not under a deadline. I was looking at the problem with fresh eyes, and the eyes saw what they saw because I was free to do it differently. The traditional way was visible. I just did not have to take it.
That was the evening I started sketching the storefront that would later become PrivaPaid, and the cryptographic shape that would later be called the macaroon. The opening chapter named the irritation. This was the day the irritation became an architecture.
The strange part was not the iframe. The strange part was where else the same shape kept showing up.
Stacker News is a forum where every post costs sats. Not metaphorically. You stake Bitcoin to submit content, and the community redistributes sats to posts that earn attention. The fee schedule is exponential. Post once, pay a small amount; post again within ten minutes, the fee multiplies by ten; again, ten times more. The cost of flooding scales exponentially. The cost of contributing thoughtfully stays flat. No verification, no gate, no question about what the contributor is. Only what the contribution was worth.
Nostr took the same idea to the social protocol. Your identity is a cryptographic keypair you generate yourself. No platform issues it, no terms of service can revoke it. The economic layer arrives through zaps: Lightning micropayments attached to notes, settled peer-to-peer, cryptographically receipted, impossible for a middleman to block or reverse. The identity is yours. The payment rail is yours. No platform owns either. A note that earns zaps has demonstrated value. A note that earns nothing has demonstrated the absence of it. The market speaks, quietly, in sats.
PrivaPaid took the same pattern to digital delivery. The macaroon gates content. The macaroon gates the comment section. When the token expires, the proof of participation expires with it. No account required for any of it.
None of the three teams had spoken to the other two. Stacker News reached the pattern from the forum problem. Nostr from the social-protocol problem. PrivaPaid from the delivery problem. Szabo had named the underlying mechanic twenty-seven years earlier in Micropayments and Mental Transaction Costs. And most of us had never read it. Four independent derivations of one architectural shape, by people who were not in conversation. When participation is not free, the filter builds itself.
That convergence is the argument. An engineer in 2026 reinvents in a macaroon library what a cryptographer named in a 1999 essay because the structure of the problem permits only one shape of answer. Show a protocol the wrong question, who are you, and it keeps answering wrong, account after account, CAPTCHA after CAPTCHA, ID upload after ID upload. Show it the right question, was your participation worth something, and the answer is the same no matter who asks it.
The Country Club and the Bar
The pattern is not exotic. Even large closed platforms are groping toward it. The most visible version charges a flat subscription fee for a verification badge and uses the payment as an input to the ranking algorithm. Paid accounts get boosted. Unpaid accounts get drowned by the feed. Not deleted, not moderated, just deprioritized. It is the same mechanism: an economic signal becomes a sort function. When the largest social networks on earth begin treating payment as a ranking signal, the pattern is no longer fringe.
But the subscription version gets the implementation wrong in two ways that are worth naming, because they are the ways any platform will get it wrong if it is not careful.
The first is the pricing model. Eight dollars a month is a rounding error in San Francisco. Less than a single coffee. It is a decision in Lagos. A meal in Bogotá. Two days of mobile data in Manila. A flat monthly fee does not filter for engagement. It filters for geography and disposable income. A farmer in Colombia with something to say about supply-chain policy, a developer in Nigeria building on the platform’s API, a student in the Philippines breaking a story the local press will not touch. Their contributions sink not because the market judged their content, but because the pricing model judged their country. That is not an incentive filter. It is a class filter with a badge on it.
A per-post economic signal is a different mechanism. The farmer does not need eight dollars a month. She stakes a few sats on the one post she cares about, and that post carries weight equal to the post of someone whose monthly subscription barely registered on the credit card statement. Not charity. Mechanism design. Skin in the game on the specific thing being said, not on the standing to say things in general.
The subscription is a country club: pay the annual fee, you are in. Per-post sats are a bar: pay for what you drink, sit where you like, nobody checks the passport.
The Filter Has to Be Readable
The second thing the subscription version gets wrong is opacity. Nobody outside the platform knows how the payment is weighted against other ranking signals. You pay, but you do not know what you bought. How much visibility does the fee buy? What other inputs compete with it? When do the rules change? The answer to all three is the same. You do not get to know. The incentive structure is buried inside a proprietary algorithm.
This is the difference between a market and a machine. In a market, the rules are legible. You see the bid, you see the ask, you see the outcome. In a black box, you pay and hope.
Stacker News publishes its fee schedule. The fee for the second post in ten minutes is ten times the first; the third is a hundred times. Anyone can read the rule, simulate it, and decide whether the cost of being heard is worth what they have to say. Nostr’s zaps are public and verifiable on the relay layer; you can audit which note received what. The macaroon logic in PrivaPaid is open source; the access-control file is in the repository, where anyone can read what got kept and what got cut. The rules are visible because the systems have nothing to hide. The filter is the price, and the price is on the menu.
The subscription-tick model inverts that. The price is on the menu, eight dollars, but the function is not. Two posts at the same price get different visibility, and only the platform knows why. That is still a filter. It has only moved the gate behind a wall the participant cannot see. An algorithm that silently drowns unpaid posts is not the absence of a gate. It is a gate with a curtain in front of it.
If the incentive structure is the filter, the filter has to be readable. Otherwise it is just an old gate with a new lock.
The Pattern Is the Pattern
The spam problem, the bot problem, the troll problem are symptoms of the same root cause: participation is free. When participation is free, the cost of noise is zero, and noise wins. Every identity gate the industry has built is a patch on top of that root cause. The patches do not fail because they are poorly engineered. They fail because they are answering the wrong question.
Charlie Munger said it for forty years. Show me the incentive, and I will show you the outcome. I had heard the line a hundred times the way one hears quotations. As decoration. Sitting with four teams converging on the same answer from four different problems, I heard it as operating instructions. The payment was the incentive signal all along. Everything around it had been overhead.
The incentive structure is the filter. The filter does not need to know who you are. It only needs to know what your participation was worth. And the rule has to be on the menu, in plain sight, where everyone can read it.
Which raises the next question. If the architecture exists, if four independent groups have already built it, if the mechanism is more legible than the systems it would replace, then why have the institutions whose job is to protect the public from opaque gatekeepers not adopted it? Why are the corrective bodies still legislating verification, still funding identity audits, still building the country club and not the bar?
The answer is the chapter that follows.