Journal · August 21, 2026
Why You Can't Pay to See Who Likes You
Five minutes on the economics of the blurred face, what the FTC's NGL case put on the record, and why we made restraint the product.
There's a small advertisement that follows me around the internet: a grid of blurred faces, a number in a red badge, and a sentence engineered to be unfinishable. Someone likes you. Tap to see who.
I run a matching service, so I've thought about that blur more than is probably healthy. Here's where I've landed: the blur isn't a preview of the product. The blur is the product. And once you see why, you understand most of what's gone wrong with how this industry treats the human heart — and why we built ours so that no amount of money can buy a clue.
The most profitable pixel in dating
An app that shows you who likes you has answered your question. An answered question stops generating revenue. An app that almost shows you can bill you indefinitely.
So the incentive is never to resolve your curiosity. It's to farm it. Admirer counts. Thumbnails blurred just past recognition. Teaser notifications that always seem to arrive in the quiet hours. Paid peeks that resolve one face and promptly present another blurred row behind it. Each mechanism converts a feeling — am I wanted? — into inventory, and the inventory only holds its value while the question stays open.
None of this requires villains. It requires nothing more than a subscription business whose core asset is your unanswered question. The design follows the incentive. It always does.
What the NGL case put on the record
Occasionally the incentive gets written down in a legal filing, and you get to see the machinery without the soft lighting.
In July 2024, the FTC and the Los Angeles District Attorney's Office filed a complaint against NGL Labs, maker of an anonymous messaging app, and its two co-founders. The complaint alleged that the app sent users fake, computer-generated messages crafted to look as if they came from real people — and then marketed a paid tier, billed at up to $9.99 a week, as the way to learn who had sent them. People who paid, the FTC alleged, didn't get names. They got "hints."
The case settled for $5 million, under an order that bans the company and its founders from offering its anonymous messaging apps to anyone under eighteen. All of it is public record; you can read it yourself.
I don't bring this up to gloat. I bring it up because it isn't an aberration — it's the teaser mechanism at full maturity. When "someone likes you" is the revenue engine, whether anyone actually likes you becomes an implementation detail.
Why we can't sell you a clue
The trustworthy way to refuse a temptation is not a policy. Policies bend the first quarter revenue misses a target. The trustworthy way is to build so the temptation has nothing to grip.
On AnonymousCrush, you name one specific person by their exact email address. We store that as a keyed cryptographic fingerprint plus a display mask — never a readable address. The person is told nothing. No notification, no count ticking upward somewhere, nothing for anyone to browse. If they independently name you too, you're both told in the same moment. If they never do, nothing ever happens — and a no never arrives, because the question itself is never delivered.
Which means we hold no teaser inventory to sell. There is no blurred face, because there are no photos or profiles anywhere in the product. There is no admirer count, because a one-sided crush produces nothing visible to anyone but the person who placed it. And there is no "someone likes you" notification — not as a policy, but as a fact you can audit. We publish every email template the system can send, all of them deliberately boring, and no such template exists.
One more refusal while we're here: we can't tell you whether any particular person has an account, and neither can anyone else. The design won't take the question. If that ever feels inconvenient, notice that it's exactly the inconvenience you'd want standing guard over you.
What money is allowed to buy here
Everything is free during early access. Eventually there will be a paid tier — Crush+, around $24.99 a year, not yet billable — and here is the entirety of what it will sell: room. Eight simultaneous crushes, against the three a free account will hold once early access ends (today, every account gets five). That is the whole feature.
It will not sell reveals. Not hints, not priority, not a boosted anything, and above all not odds: the matching engine treats a paid account and a free one identically, and it always will. Selling better chances would mean selling other people's attention — which was never ours to sell.
The same reasoning sits behind a smaller decision. You choose how long each crush stays sealed — a month, three, six, or a year — and renewing it is free on every plan: one quiet reminder before it lapses, one tap to continue. We will never charge rent on a hope. A business that does has rebuilt the teaser incentive, just slower and with better manners.
The cap no one can buy past
Every account carries a hard ceiling of ten active crushes. Crush+ will not raise it. No future tier will. It is not for sale at any price.
A safety limit that vanishes when you pay isn't a safety feature; it's a price point. The cap exists because this is a tool for a few real, specific hopes, not a dragnet — and if a limit is worth having at all, it's worth holding against our best-paying customer most of all. There's more on this thinking in our safety design.
Restraint is the product
So what do we actually sell? A silence that holds. Even after two people match, consent stays doubled: identity is revealed field by field, only as fast as the other person agrees, and messaging opens only if both allow it. And if someone steps back after matching, the other side sees the same quiet ending an expiry would have produced. No one learns who closed the door.
If you never touch our product, take the transferable lesson anyway: whenever an app shows you a blurred face, ask who profits if the blur stays blurred. The answer tells you whether you're looking at a matchmaker or a slot machine.
And if you'd rather no service ever hold a question about you at all, you don't need an account to say so — anyone can permanently opt out, and the system will never match or email you again. Of everything we've built, that page — the one that sells nothing, to people who owe us nothing — may be the one I'm proudest of.
But if there is one person you've been meaning to ask — quietly, with no audience possible and nothing lost by asking — that is exactly what we built.
