ChatTJB, a chatbot with no model behind it, has logged over 30,000 queries as of August 6, with peak traffic hitting 5,000 prompts per hour. The interface looks like ChatGPT or Claude. The responses come from Tucker Bryant, a 32-year-old former Google project manager, typing and sketching answers by hand in San Francisco.
The site markets itself as "the leading chat interface powered by AI" — with AI redefined as "average individual." Ask ChatTJB how to fry an egg, Bryant writes back. Ask for an image of a skunk wearing sunglasses, he scribbles one. Response times are not fast. The site currently notes that "thousands of people are writing in right now."
Bryant launched the project in April but it did not attract attention until July 27, when he paid $6,000 to put a billboard up in San Francisco. Coverage followed, and the queue exploded. The billboard is scheduled to come down in late August.
Key facts
- 01ChatTJB has logged over 30,000 queries as of August 6, with a peak of 5,000 prompts per hour.
- 02Creator Tucker Bryant, 32, a former Google project manager, answers every prompt himself — no LLM.
- 03A $6,000 San Francisco billboard erected July 27 drove the traffic spike; it comes down in late August.
- 04Over 3,000 people applied to help; Bryant enlisted 10 volunteers.
- 05ChatTJB Pro, a $5-a-month tier Bryant calls 'a worthless tier,' has four paying subscribers.
Bryant calls ChatTJB an art project first. He vibe-coded the site using Lovable, so he is not an AI refusenik. His concern is what he calls cognitive surrender — the reflex to hand small human decisions to a language model. He cites a moment when he asked an LLM whether to wear short or long sleeves on a 65-degree day in San Francisco, a city he has lived in for seven years.
The prompts he receives split roughly between jokes and something more earnest. Recent drawings include a frog eating spaghetti, San Francisco mayor David Lurie as Godzilla, and two black cats eating grass on a table. Written questions range from what to make for dinner with the contents of a fridge to what to do for a birthday in two weeks when the user does not want a party.
The shift from prank to confession happened quickly enough that Bryant noticed the pattern.
The economics do not work as a business, and that is by design. ChatTJB Pro, a $5-a-month tier that Bryant openly labels "a worthless tier" on the site, has four subscribers. That is $20 a month against a $6,000 billboard. He has recruited 10 volunteers to help answer prompts, selected from more than 3,000 applicants.
The read on why users keep coming back is not that Bryant gives better advice than a frontier model. On YouTube, he described ChatTJB as "the worst LLM of all time." The draw appears to be something else.
ChatTJB works as commentary because the industry it satirizes has spent 2026 racing in the opposite direction. OpenAI, Anthropic, and Google have shipped increasingly agentic products this year designed to remove the human from more steps of more decisions. ChatTJB inverts the pitch: the human is the entire product, latency and all. It is a small piece of counter-programming against a market where the working assumption is that any task a person does today, a model should do tomorrow. Whether the joke has legs past late August — when the billboard comes down and the traffic likely follows — is the real test of whether cognitive surrender is a thesis people will pay attention to, or just a bit.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
