Ex-Google Employee Gets Flooded With Submissions For ‘ChatTJB’ Bot Powered By Just Him

Tucker Bryant, a Stanford graduate and former Google product manager, created ChatTJB, a website that mimics an AI chatbot but is powered solely by Bryant himself, and after a billboard in SoMa last month, submissions surged to roughly 5,000 per hour, forcing him to pause the service and consider a community-driven approach. Bryant has received more than 1,000 messages from volunteers interested in becoming human 'AIs' and is cautiously testing moderation tools, telling Futurism, 'I’m going to move slowly. This was just intended to be an art project.'

After launching a billboard in SoMa last month, Tucker Bryant, creator and sole operator of the site ChatTJB, which is powered by the “average individual,” began receiving thousands of requests from users seeking advice and volunteers eager to become human “AIs.” The project started as an art project by Tucker Bryant, a Stanford graduate and former Google product manager who wanted to poke fun at the increasingly automated world of AI chatbots, as SF Gazetteer reported https://sf.gazetteer.co/the-new-ai-billboard-locals-actually-like last week. He launched ChatTJB https://chattjb.org/ in April as a chatbot-style website where every response would come from Bryant himself, rather than an artificial intelligence. Rather than generating answers with a large language model, ChatTJB sends each question to Bryant, who reads it and responds himself, when he is "awake and motivated." The site's popularity began to skyrocket after Bryant put up a billboard in SoMa last month near Sixth and Shipley Streets. Photos of the sign began circulating online, including on Reddit https://www.reddit.com/r/sanfrancisco/comments/1vb6gkb/parody ai billboard/ , sending the number of questions submitted to ChatTJB from a few dozen to thousands, at one point reaching roughly 5,000 submissions an hour. Bryant told Futurism https://futurism.com/artificial-intelligence/guy-chatbot-manually-responding-chatjtb he was personally answering them all. He said the idea came partly from the concept of “cognitive surrender,” which he describes in a YouTube reel https://www.youtube.com/shorts/hXDplHV9KGA as the public's willingness to trust AI-generated advice, even when it's wrong. He said he had recognized the tendency in his own interactions with chatbots and wanted to create an experience that looked like a conventional AI service but required users to contend with the unpredictability and inconvenience of dealing with an actual person. The approach can result in wildly different responses depending on the question. Bryant has even produced images for users, although his version of image generation involves hastily drawn graphics https://www.instagram.com/p/DXw70xnhFLE/ rather than an AI image model. According to Futurism, while plenty of people have used ChatTJB to troll him, Bryant said others have come to the site with surprisingly personal questions, prompting him to drop his usual casual approach and respond more like a supportive friend. Some users have reportedly reminded him to eat, drink water, and take breaks as Bryant works through the enormous backlog. He notes that the concerned messages from users have been another reminder of the human element that’s been built into the project. The flood of questions eventually forced Bryant to pause the service, and he's now cautiously experimenting with turning ChatTJB into a community effort. He's reportedly received more than 1,000 messages from people interested in being “average individuals,” and he's begun allowing some approved volunteers to respond to questions while he tests moderation tools. Bryant said he wants to proceed cautiously, as he does not want to put users or volunteers at risk. “I’m going to move slowly,” Bryant told Futurism. “This was just intended to be an art project.” He said the response has been meaningful enough that he is willing to see where the community version goes, but that ChatTJB could be scaled back or paused again if the experiment does not work. Related: Fake, Satirical Startup 'Replacement.AI' Puts Up Haunting Billboards In SF, NYC https://sfist.com/2025/10/27/fake-satirical-startup-replacement-ai-puts-up-haunting-billboards-in-sfs-castro-nyc/ Image: gluggin/Reddit