Inside ChatTJB, the Viral “AI” Chatbot That Is Really Just One Guy

ChatTJB, a San Francisco chatbot billed on a billboard as an AI-powered interface, is actually one man answering prompts by hand — and that stunt has turned into a widely discussed art project with tens of thousands of messages. Its creator, former Google project manager Tucker Bryant, says the joke is meant to make people rethink how casually they hand over judgment to software.

Bryant launched the project in April, but it exploded after a $6,000 billboard went up in San Francisco on July 27, drawing more than 30,000 queries by August 6 and reaching peaks of roughly 5,000 prompts an hour, according to Bryant. The project has become part performance art, part commentary on the growing habit of outsourcing everyday decisions to chatbots.

The premise is simple, and deliberately absurd: instead of a large language model, ChatTJB is powered by an “average individual.” Bryant sits down and replies himself, sometimes with doodles, sometimes with advice, and sometimes not at all. The result is a viral experiment that lands somewhere between comedy, confession booth and a live demonstration of the human cost of always asking a machine first.

What is ChatTJB and why did it go viral?

ChatTJB is a chatbot in name only. Behind the interface is Bryant, a 32-year-old public speaker, artist and former Google employee, who personally reads and answers incoming prompts rather than relying on AI-generated text.

The project caught fire after the billboard appeared in San Francisco. The sign describes ChatTJB as a “leading chat interface powered by AI,” but the twist is that the letters stand for “average individual.” That joke, paired with the visual punch of a prominent city billboard, helped turn a niche art concept into a local and then national curiosity.

According to Bryant, the service had existed in a quieter form for months, but the billboard changed everything. The attention brought a flood of messages, a wave of social sharing, and a pace of demand that would be difficult for even a well-funded startup to sustain — especially when the product is one person typing replies by hand.

How the project works

ChatTJB’s mechanism is intentionally low-tech. Users submit a question, Bryant sees it, and he answers. If someone asks for a drawing of a skunk in sunglasses, he sketches it himself. If another person wants help figuring out dinner, he may offer a suggestion. If he is not available, the message may simply wait.

The website itself leans into the joke. Bryant used a no-code, AI-assisted web builder to assemble the site, but he has made clear that the “model” inside is not a model at all. He has even joked publicly that ChatTJB is the “worst LLM of all time.” The humor is part of the product, but so is the point: users are not talking to an optimized machine, they are talking to a single person with limited time, taste and attention.

Why does Bryant see it as an art project?

Because Bryant says the central purpose is commentary, not utility. He describes ChatTJB as first and foremost an art project, a way to respond to a cultural moment in which people increasingly ask software for advice that used to come from friends, family or personal judgment.

In his view, the project sits inside a larger social shift: people are becoming comfortable treating chatbots not just as tools, but as extensions of their own thinking. Bryant says that makes the work relevant as a cultural mirror, even if the premise is intentionally ridiculous.

Bryant says he sees the project as a way to comment on a strange period in which people routinely defer to AI for decisions they could make themselves.

That concern is not abstract. Bryant says the idea grew out of his own experience using large language models for trivial questions, including whether he should wear short sleeves or long sleeves on a 65-degree day in San Francisco. He says that kind of interaction made him reflect on how quickly convenience can turn into dependence.

His discomfort centers on what he calls cognitive surrender: the gradual habit of accepting machine-generated answers with less and less scrutiny. ChatTJB turns that fear into a live joke. Instead of asking an algorithm, people ask a person and often discover that what they wanted was not certainty, but permission to think aloud.

What kinds of questions are people asking?

People are asking the kinds of questions that reveal how intimate chatbot use has become. Some prompts are playful, such as requests for drawings or silly scenarios. Others are ordinary life choices: what to eat, how to use ingredients already in the fridge, or whether to throw a birthday party two weeks from now.

Bryant says many users initially approach the project as a gag. But once they realize there is a real human on the other side, the tone changes. Questions become more personal. People open up about frustration, uncertainty and low-stakes decisions that still feel emotionally loaded.

From jokes to genuine advice

That shift is one reason ChatTJB has resonated more than a one-off prank might. The joke lands immediately, but the replies can become unexpectedly tender.

Bryant says the appeal may lie less in the quality of the advice than in the simple fact that a person is taking the time to respond. In his view, many users do not need expert guidance; they want a human sounding board, even if that human is unqualified, imperfect and occasionally slow.

Bryant says people seem to value the chance to bounce ideas off another person rather than receive polished, expert-level guidance.

The requests he cites suggest a familiar modern tension. Many of the most mundane choices in life are now framed as prompts for optimization: dinner planning, scheduling, wardrobe decisions, social dilemmas. ChatTJB strips away the illusion of machine omniscience and replaces it with something older and messier — human judgment.

How big has ChatTJB become?

The scale surprised Bryant. He expected the project to last only a few days, but momentum kept building after the billboard appeared and media attention followed. By Bryant’s count, ChatTJB had passed 30,000 queries within about 10 days of the billboard going live.

At its busiest point, he says the system was receiving around 5,000 prompts each hour. For a one-person project, that volume is significant not only because of the number, but because it highlights the mismatch between internet virality and human capacity. A chatbot backed by actual AI infrastructure can absorb surges. A person cannot.

To keep things manageable, Bryant has recruited 10 volunteers, but even that small team was highly selective. More than 3,000 people applied, he says, underscoring both the unexpected demand and the desire among many users to participate in the experiment, not merely consume it.

Volunteer interest and public curiosity

The volunteer response suggests that the project has evolved beyond novelty. Some people want to help moderate, draw, or answer prompts. Others seem drawn by the chance to be part of a cultural moment that pokes fun at the AI economy while still using its aesthetics and language.

The project also taps into a wider fatigue with polished tech narratives. In an era crowded with startups promising to automate every kind of work and every kind of thought, a deliberately limited, human-only chatbot can feel both refreshing and subversive.

What is ChatTJB Pro?

ChatTJB Pro is a paid support tier for people who want to help fund the project. Bryant says the tier is intentionally useless: subscribers do not receive extra features, special access or premium responses. They are simply supporting the art.

So far, he says, only about four people have signed up at $5 per month, which means the project has not recovered the cost of the billboard. The ad space, which Bryant paid for at roughly $6,000, is set to stay visible until late August.

That financial mismatch appears to be the point. Bryant says he did not build the project to generate revenue or scale into a business. Instead, he wants it to function as a temporary public intervention — a strange, mildly absurd thing that interrupts the usual flow of tech promotion.

Why the money matters anyway

Even a project that is not meant to be profitable still reveals something about how audiences assign value. People are willing to pay for novelty, access and identity signaling. In this case, they are paying for the knowledge that they helped sustain a joke with a message.

That may be one reason the project has attracted attention beyond the usual tech-news cycle. It has a physical billboard, a viral concept and an anti-optimization aesthetic, but it also offers a subtle reminder that the economics of attention are often stranger than the products themselves.

How does ChatTJB fit into the current AI debate?

ChatTJB lands at a moment when AI products are everywhere and skepticism is rising. Consumers are being told that chatbots can draft emails, plan trips, tutor children, summarize documents and even make decisions. At the same time, critics warn about overreliance, hallucinations, privacy risks and the danger of ceding too much judgment to systems that may be fluent but not reliable.

Bryant’s project lands squarely in that tension. It uses the look and feel of a modern AI service to show how quickly people trust the interface — even when the promise underneath is obviously false. That makes the project funny, but also useful as a kind of social experiment.

There is an especially sharp irony in the fact that Bryant is not anti-technology. He used an AI-assisted tool to build the website, and he speaks in a measured way about the value of tech. His critique is narrower: he is warning against treating every convenience as if it were wisdom.

A critique without rejection

Unlike a Luddite protest, ChatTJB does not reject AI outright. It borrows AI’s form, language and cultural cachet, then flips the premise. That allows Bryant to make his point without pretending the technology has no value.

The result is more self-aware than hostile. It asks whether people are reaching for AI because it truly improves a decision, or because it has become the easiest default. In that sense, the project is less about machines than about habits.

Timeline of ChatTJB’s rise

The project’s rapid rise helps explain why it has punched above its weight.

What makes people keep coming back?

Part of the answer is novelty. Part is humor. But a deeper reason may be that ChatTJB offers something mainstream AI products often do not: visible humanity.

Even when the replies are imperfect, users know a person is sitting with the question. That knowledge changes the emotional texture of the exchange. The answer is not optimized, but it is accountable in a way that machine-generated text often is not.

There is also a kind of relief in the project’s honesty. In a market full of exaggerated claims, ChatTJB says exactly what it is, even while the billboard appears to say the opposite. The punchline is not deception for profit; it is deception as critique.

For some users, that seems to be enough. They are not looking for a better assistant. They are looking for a stranger with enough attention to react to their question. In a digital environment that often feels automated, that can feel surprisingly meaningful.

What the project says about the future of AI

ChatTJB suggests that the future of AI may be shaped not only by better models, but by growing skepticism toward the idea that every question needs a machine answer. The project’s success hints at a public appetite for alternatives, satire and a reminder that human judgment still has value.

It also points to a likely split in how people relate to AI. Some will continue to seek speed, scale and precision from software. Others may become more interested in friction, context and conversation — qualities that are harder to automate and, at times, more reassuring.

Bryant’s experiment does not resolve that debate. But it does give it form. By making a chatbot out of a single human being, he exposes both the absurdity and the appeal of our current dependence on digital advice.

And that, ultimately, may be why ChatTJB has struck a nerve. It is funny because it is fake, but also revealing because the fake is close enough to the real that people keep using it anyway.

Bryant says he wants the project to encourage people to reconsider their relationship with technology and to make an anxious moment feel a little lighter.

That may be the most accurate description of ChatTJB: a temporary, human-sized counterweight to the promise that AI can answer everything.