AI Labs Court the Creator Who Trolls Them
Original: He Trolls AI Labs for a Living. Now They All Want to Work With Him
Why This Matters
Viral criticism from outsiders is now shaping how AI labs manage public perception.
Matt Husk, a comedian with 2M followers who exposes AI chatbot failures, says OpenAI, Anthropic, and xAI have all approached him to collaborate after his deadpan trolling videos went viral.
Matt Husk started making videos on a whim while working a landscaping job, asking ChatGPT absurd questions and recording its failures. A year later, he has 2 million combined followers on TikTok, Instagram, and YouTube, a deal with talent agency Unicorn, and three of the biggest AI companies knocking on his door.
Speaking at WIRED World Fair in Miami, Husk told interviewer Evan Ratliff that OpenAI, Anthropic, and xAI have all inquired about working with him. xAI reportedly offered him a Tesla for a week in exchange for exclusive content. OpenAI's outreach was less generous — Husk said the company made an offer so low he flatly refused, then tried sending gifts and asked for his home address. He turned that down too.
OpenAI also offered early access to a fix for its broken timer — a bug Sam Altman publicly acknowledged on the podcast Mostly Human after being shown one of Husk's videos — if Husk shared his account. He declined that as well.
His most popular videos show ChatGPT's voice model botching simple tasks: counting to 100, beatboxing, timing a mile run. In one clip, Husk asks it to negotiate Craigslist bread down from $10,000 to $5. The model settles at $400. His first viral video came when he told ChatGPT a nuclear bomb had gone off and received the response: "I'm rooting for you."