The Accelerationists Are Trying to Stop Anthropic

The Accelerationists Are Trying to Stop Anthropic

Anthropic has done the impossible. It has united the accelerationists, safetyists, effective altruists, and ethicists around one cause. Unfortunately for the company, that cause is stopping Anthropic. Many prominent figures from different factions of the AI landscape have signed on to “The MisAnthropic Statement,” criticizing Anthropic as a company that feigns concern for AI safety while plowing forward in the name of profit, just ahead of the company’s planned initial public offering. “Anthropic warns that recursively self-improving AI could concentrate power and escape human control, yet races to build it first,” the statement reads. “The company uses safety branding to court investors, employees, and the public—even as it has quietly weakened its safety pledges and lobbied for self-serving regulation. Anthropic is not a safe AI company, and we must reject its reckless attempt to amass power in the name of the greater good.” The list of signatories is a Who’s Who of the AI safety debate, in part because the statement has been kept so vague that just about anyone can feel comfortable putting their name on it—which is how you end up with people like Émile Torres, perhaps the most prominent critic of the accelerationist movement, putting their name alongside Guillaume Verdon aka Beff Jezo, the founder of that very movement. Somehow, despite the letter being written in a way to make it maximally agreeable, some found it to be too harsh. Open-weights advocate and Trillium Labs founder Nathan Lambert, for instance, said on Twitter that while he does believe “Anthropic is power seeking” and doesn’t acknowledge the risks it creates, he opted not to sign because he “found this statement too strong for my taste.” Dan Hendrycks, Director of the Center for AI Safety, noted on Twitter that the statement was also strong enough to scare off some folks who didn’t want to ruin their relationship with Anthropic (and, probably more importantly, mess up their own money). “Numerous AI safety groups told us privately they agreed with this statement, yet felt unable to sign without risking relationships or future funding,” he wrote. “Anthropic has almost, but not entirely, captured the ecosystem meant to hold it accountable.” Hendrycks has a history of uniting people around very agreeable and ultimately meaningless statements. His organization previously secured the signatures of OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei to sign on to a statement on AI extinction risk that read, “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” Seems reasonable enough, but here we are with Hendrycks rallying others to sign on to another statement because Anthropic is apparently not only failing to abide by the last one but also weaponizing it for its own benefit. Surely, a solid share of the signatories are genuine in the concern they express and believe Anthropic to be misleading the public about just how invested in AI safety it truly is. Gary Marcus, an AI researcher and noted skeptic of the potential of LLM models to achieve superintelligence, signed the statement and expressed on Twitter that his concerns are not isolated to Anthropic. He was joined by dozens of other researchers and academics at institutions like Harvard, MIT, and UCLA who are all warning that Anthropic is not the safety-first lab that it claims to be. But some of the signatures should probably be accompanied by an asterisk. Take Igor Babuschkin, co-founder of xAI. While his fellow founder, Elon Musk, is not listed, the statement borrows Musk’s favorite insult to hurl at Anthropic—”Misanthropic”—for its title. Babuschkin left xAI in August 2025 to start an investment firm focused on AI safety research, but his former company remains deeply entangled with Anthropic’s business. SpaceX, which owns xAI, is currently renting its Colossus data center to Anthropic to train Claude, which doesn’t seem like the kind of thing that a company concerned about the development of a dangerous superintelligence would do. Then there’s Verdon, better known on Twitter as Beff Jezos. He’s the leader of the all-gas, no-brakes view of AI who believes that we should push for the rapid growth of technology and capitalism at the expense of everything else, essentially because developing the technology will, in his view, solve all the problems created in the process. He also believes that humans are not essential to consciousness and he “has no particular allegiance to the biological substrate.” Verdon has shown little enthusiasm for slowing AI development in the name of safety, but he has decided that Anthropic is behind the ousting of Jensen Huang at Nvidia—an accusation unsupported by public evidence—and is now concerned about its accumulation of power. His concern is not with Anthropic skirting the safety practices that it advocates for but the potential for the company to lock others into safety requirements with regulatory capture. It ceases to be a race to extinction if you ban others from the track, you see. It’s unlikely that this letter will do anything to stop or even put a dent in Anthropic’s IPO. But it is at least interesting to see just how many factions in the AI space do want the company slowed, even if it is for entirely different (and usually self-serving) reasons.

Original Source

Read the full article at Gizmodo →

KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.