Protesters Gather on Anthropic IPO Day
Around 100 people protested in the rain outside Anthropic's San Francisco office on Thursday, the same day the AI lab held the largest initial public offering in history. Their main demand was a ban on recursive self-improvement (RSI), a technology where advanced AI models autonomously create newer versions of themselves.
Extinction Fears Drive Protest
Protesters chanted "Ban RSI, we don't want to die" and held signs reading "Don't invest in extinction." They argue that RSI could hasten doomsday scenarios and eliminate human life.
"These people are going to make a profit off of a technology that even they admit has a chance of eliminating life on Earth," said Tarnel Abbott, a 72-year-old retired librarian.
Key Voices
- Jacob Coxon, a former Anthropic researcher who resigned in September with safety warnings, said only a few places worldwide can develop RSI in 2027 and Anthropic's office is one. He added that Silicon Valley insiders privately admit they don't want RSI developed soon.
- Emily Tan, 31, who travelled from Los Angeles, called the situation a "black box" and said most people don't know what's happening.
- Tyler Tracy, 26, a researcher at Redwood Research, questioned whether Anthropic can prioritize safety under shareholder pressure.
- Daniel Kokotajlo, a former OpenAI employee now heading the AI Futures Project, rejected the argument that Anthropic and OpenAI are "the responsible good guys" racing against rivals.
Anthropic's Response
The company stated it has not yet reached the RSI stage. Founded by former OpenAI employees who vowed to prioritize safety, Anthropic faces scrutiny over whether it can maintain those values after its public listing.
Open Letter and Counter-Protest
A day earlier, an open letter signed by researchers, entrepreneurs, and academics accused Anthropic of not being a safe AI company. A single counter-protester, Barak Gila, 31, thanked Anthropic for the technology and local economic impact, leading to a brief confrontation.
Many protesters acknowledged using AI tools but feel development is advancing too quickly without sufficient safeguards.