Art project flips AI model on its head with real human responses

Former Google engineer Tucker Bryant created ChatTJB, an experimental project that playfully subverts the concept of AI assistants. Superficially, it resembles a standard chatbot with a minimalist interface, but unlike any language model, responses come from real people. Bryant originally conceived the project as a critique of growing AI reliance and people's tendency to outsource important decisions to machines. After spending $6,000 on a San Francisco billboard advertisement, the site received over 100,000 questions, with the volunteer waitlist growing by roughly 1,000 people daily. Bryant initially responded to questions personally, but demand ballooned to 5,000 questions per hour, forcing him to rely on volunteers to keep the platform running. Though it started as artistic satire, the project gained genuine traction, and Bryant is now open to partnerships to make it sustainable. The initiative reveals a peculiar cultural tension: while companies race to make machines seem more human, there is real interest in a space where people respond like machines.