Talk to this twenty-year-old founder for five minutes and you stop betting against him. He built the tool that tells you whether what you’re looking at is real, the one that recently caught a fake mugshot in under two minutes, and he’s more certain about where it’s going than anyone has a right to be. So far, he keeps being right.
Sam Cons is the twenty-year-old founder who’s spent the better part of the last year on a question most of the technology industry would rather not sit with: how do you know that what you’re looking at is real? He built a company around the answer. More than five million times now, someone’s handed that company a photo, a headline, a screenshot, and asked it exactly that, then waited for the verdict. The company is Cytation AI. He started it in Canada, took no outside money, and owns every share of what he’s made.
At first he comes across like a lot of young founders, friendly, composed, easy to underestimate. Then the subject turns to Cytation, and the impression doesn’t survive it. He locks in so completely he seems to forget he’s being interviewed at all, and the easy manner narrows into something more deliberate, harder to look away from. He doesn’t pitch and he doesn’t hedge. He talks about where Cytation is going with the specificity and certainty most people reserve for things that have already happened, not as a hope he’s chasing but as a destination he can already see, and then he walks you through exactly what stands between here and there. The vision runs years ahead of where the product sits today. What keeps it from tipping into bravado is that he treats the distance between what Cytation is and what he’s certain it will become as a problem of execution rather than belief. He moves like someone who’s already decided how this ends and is simply working backward toward it, faster than seems reasonable.
I saw what that looks like the first time I sat down with him, about five weeks ago. Cytation had roughly 125,000 users. He told me, flatly, that it would reach 140,000 by the end of June. Not hoping. Stating. I told him what you tell someone who says things like that: aim big, but be happy with wherever it lands. On June 17th it crossed 140,000, on the number.
That same intensity shows up in what he wants out of a conversation. Most founders, pressed on how a product is landing, are listening for whether you’d use it. He’s listening for the opposite. Partway through, he asked me something no founder trying to be liked would ask: what Cytation would have to get wrong, just once, for me to never trust it again. It’s the question of someone who understands that a trust product doesn’t survive a single bad answer, and who would rather find that failure himself, early, than have it found for him later. He works a conversation the way he works the product, hunting for what’s broken and closing in on it before anyone else notices it’s there.
I’d come to all of this sideways. I first paid attention to Cytation about three months ago, and not as a reporter; it ended a small crisis I was standing next to. A friend called me one afternoon in a panic. He’d been scrolling and stopped cold on his brother’s face in a mugshot, a name and a charge printed plainly underneath. His brother had never been arrested. He’d called him twice and gotten nothing, not knowing his brother was simply at work with his phone in his pocket. I’d been hearing about Cytation for weeks, so I told him to try it. He ran the photo: AI-generated. He ran the claim: false. A beat later his brother called back, mildly annoyed, asking why he’d called twice in the middle of a workday. He’d never seen the photo. He didn’t know he’d been arrested, because he hadn’t been. The whole thing lasted about two minutes, and the lie never reached the person it was about. It detonated in someone who loved him instead.
I followed the company from that afternoon on, and eventually it put me across a table from its founder, where the thing I most wanted to understand was the gap his product had closed in those two minutes. That distance, between the moment a lie reaches you and the moment you can prove it’s one, is exactly what Cons set out to close. The idea arrived the way the good ones usually do, out of something that bothered him and wouldn’t leave. He’d watched a public figure get redefined in real time by the internet, not through one big lie but through a thousand small clips lifted out of context, each technically real, stitched into a portrait that was completely false. He watched a consensus harden into something people treated as truth, with no friction anywhere, no moment where anyone had to stop and ask whether the story actually held. “I kept thinking, how did this happen,” he says. “And then, how do I fix it.” The answer was almost embarrassingly simple. Build a system that tells you whether something is true or false.
The idea, he’s quick to say, was one percent of it. “Executing it was hell.” What followed was more than a hundred thousand iterations, more than eight million data points trained into the detection model, and a working rhythm he sums up in four words: ship, break, retrain, repeat. Build a version. Watch it fail on something it should have caught. Retrain it. Ship it again the same day. Founders tend to romanticize this part in hindsight. He’s still in it, and he talks about it the way other people talk about something they love and resent in equal measure.
What’s striking isn’t only the effort, it’s the direction he aimed it. For Cons the reasoning runs on two tracks at once, and he doesn’t pretend otherwise. One is close to moral. Truth itself felt like it was coming loose, and he couldn’t see anyone building the thing that might hold it together while the rest of the industry built the thing pulling it apart. The other is plainly commercial. A massive, worsening problem with no real solution isn’t a cause, it’s an opportunity. He saw both at once, and he doesn’t feel the tension other people expect him to.
That same independence of mind has a financial counterpart: he’s kept the company entirely his own, and his reasons stack. A company whose only product is trust, he argues, can’t afford an investor quietly pulling it off-mission. He wanted the demand proven before any money entered the picture, so the growth meant something. And underneath both, something simpler: he never felt he needed permission, or a runway, to build something he already believed in. Independence, in his telling, isn’t a financing decision. It’s part of the product.
Kept lean and undiluted, the company grew quietly for a while, the way useful tools do. Then he shipped the thing he’d seemingly been building toward all along, and it gave away how big he was actually thinking. The Feed pulls posts from across the major platforms, Instagram, X, TikTok, Reddit, Facebook, YouTube, the news, and runs each one through verification before it reaches you, stamping it with a real-time Truth Score. It gathers all of social media into a single stream you can scroll, every post already checked by the time you land on it. He hadn’t set out to build something people grab in a panic and forget, the way you grab a fire extinguisher. He’d set out to make verification a place people live in rather than an emergency they visit. The numbers moved the way they move when a founder turns out to be right: after the Feed launched, people who used to open Cytation once and leave started opening it daily, retention jumped roughly tenfold, and the growth curve went nearly vertical.
Where it goes from here he describes in deliberately flat terms, which from him reads less like modesty than like someone refusing to jinx it. A verified internet, where being able to tell what’s real is simply part of the infrastructure, there by default rather than something you go hunting for. A Feed that becomes the place people go for information they can believe. And, further out, the enterprises and platforms and institutions that will need verification they didn’t build themselves and can’t quietly compromise. He won’t dress any of it up. “We’re only getting started” is about as far as he’ll go.
I keep coming back to my friend’s two minutes. His brother was never arrested. For about two minutes, the internet said otherwise, and there was no way to argue with it. Then, very quietly, there was.
The thing Sam Cons is building isn’t loud, and neither is he. It is, at bottom, a single question aimed at the moment just before you forward something, share something, believe something. The same question the whole company is built on, the one he keeps turning back on the people around him, including, you suspect, himself.
What are you so sure about?
