SAEDNEWS: A new chatbot advertised on a San Francisco billboard looks like another cutting-edge AI service, but there is a surprising twist: every response is written by a single human, artist and former Google employee Tucker Bryant.
According to SaedNews: The next time a chatbot gives you a smooth, confident answer, you might want to stop for a moment and ask a simple question: Who—or what—is actually answering me?
That is precisely the question behind ChatTJB, an unusual new project being promoted on a billboard in San Francisco. At first glance, the advertisement appears to be promoting yet another advanced artificial intelligence service. In large letters, it describes ChatTJB as a leading interface for an AI-powered chatbot.
But there is a catch.
A small disclaimer near the bottom of the billboard reveals that the “AI” in this case does not stand for Artificial Intelligence. Instead, it refers to “Average Individual”—an ordinary person.
And that person is Tucker Bryant.
Bryant, an American artist and former Google employee, is the human being sitting behind ChatTJB. Rather than connecting users to a large language model capable of producing answers in seconds, the service sends their questions to Bryant, who reads them, thinks about them and eventually writes a response himself.
The concept is intentionally absurd.
At a time when technology companies are competing to make AI systems faster, more convincing and increasingly capable, Bryant has created a chatbot that does almost the opposite. It is slow by design. Its answers may be uncertain. They may even be wrong. And instead of generating images with sophisticated software, the system can respond to image requests with drawings made by hand.
Bryant has jokingly described the project as the “worst large language model in history,” while also suggesting that it could change the way people interact with AI.
The website behind ChatTJB describes the service as “handcrafted intelligence, made by a human.” Rather than an LLM, it presents the idea of a “Large Language Experience” operated by a single person.
Its explanation takes the joke even further. Instead of having questions processed by a statistical model, each request is handled by what the project calls a dedicated “biological reasoning system.” In practical terms, that means there is no mysterious artificial system deciding what to say.
There is simply Bryant.
A user asks a question. Bryant reads it. He thinks about it. Then, whenever he is awake and motivated, he writes the answer.
That deliberately inconvenient process is the heart of the project.
Behind the humor, Bryant says there is a more serious idea at work: “cognitive surrender.”
The term refers to a tendency to hand over critical thinking to chatbots, particularly when their responses sound confident, polished and authoritative. Even when a chatbot is wrong or misleading, its fluent presentation can make an answer feel more trustworthy than it actually is.
That is the behavior Bryant wants people to notice.
He argues that users have become accustomed to receiving instant answers that sound intelligent and certain. As a result, some people may stop questioning what they are being told the moment they enter an AI interface.
Bryant is not presenting himself as an opponent of artificial intelligence. His criticism is aimed at what can happen when people become too comfortable with it.
The distinction matters.
A chatbot does not have to be correct simply because its answer sounds sophisticated. Fluency is not the same thing as truth, and confidence is not proof.
ChatTJB tries to make that lesson impossible to ignore by making the chatbot experience intentionally frustrating.
Bryant says ChatTJB is designed to be “infinitely worse” than today's mainstream chatbots.
Its responses will be slow. They probably will not be correct. They might be correct, but users will not necessarily know. They will certainly be ambiguous at times.
In other words, users are not supposed to have a seamless experience.
That is the point.
Imagine asking a conventional chatbot an important question and receiving a polished response within seconds. Now imagine asking ChatTJB the same question and waiting for one person to read it, consider it and manually respond.
The contrast exposes something that is easy to overlook when using modern AI: just how much we have begun to associate speed, confidence and polished language with intelligence.
Bryant wants users to remember that feeling the next time a language model gives them an answer that sounds perfectly reasonable.
His most striking example involves a financial or tax question. Instead of confidently explaining the issue, ChatTJB might respond with a haiku about crows.
It sounds ridiculous—and deliberately so.
The hope is that after encountering such an obviously unsuitable answer, users will become slightly more skeptical the next time an AI system provides an impressively fluent response to a serious question.
The experiment is therefore less about replacing AI than about changing the user's reaction to it.
There is also something unusual about the transparency of the project.
Traditional chatbots often feel almost human because the machinery behind the interface is hidden. ChatTJB reverses that relationship. The “intelligence” behind the interface is explicitly a person.
There is no illusion of an enormous computational system operating somewhere in a data center. There is one individual responsible for the response.
That makes the experiment both funny and revealing.
Bryant's project turns the idea of a chatbot inside out. Instead of asking how close a machine can come to behaving like a human, it asks what happens when a human deliberately behaves like a chatbot—and does a poor job of it.
The result is a strange technological parody with a serious warning at its center.
As AI becomes increasingly embedded in everyday life, users will encounter more answers that are fast, articulate and persuasive. The harder challenge may not be learning how to use these systems, but remembering when not to surrender our own judgment to them.
That is the lesson Bryant hopes ChatTJB will leave behind.
And perhaps that is the project's biggest irony: the chatbot that does the least to imitate artificial intelligence may end up encouraging people to think more carefully about the technology that does it best.
Bryant's final conclusion is deliberately provocative: the next stage in the evolution of AI could be “human meaninglessness.” Whether that idea is satire, criticism or something in between is left largely to the audience.
But one thing is clear. ChatTJB is not trying to convince people that a machine is human.
It is asking humans to remember that they are still responsible for thinking for themselves.
