← Wednesday, November 18 deck

09Safety2 sources posting

CAIS Posts the MisAnthropic Statement, a Coalition Critique Ahead of Anthropic's IPO

Ahead of Anthropic’s IPO, a rare coalition across AI safety, open source, e/acc, science, and law has signed the MisAnthropic Statement. Read the statement or sign at:

Context

Why it matters

A public statement signed across usual rival camps, posted by a safety institute on the eve of a huge listing, is a coordinated attempt to shape how Anthropic is seen before it goes public.

The Center for AI Safety posted that, ahead of Anthropic's IPO, a rare coalition across AI safety, open source, e/acc, science, and law has signed the MisAnthropic Statement, and pointed readers to a site to read or sign it. A later post from MTS named signers including Hugging Face cofounder Thomas Wolf, xAI cofounder Igor Babuschkin, MIRI President Nate Soares, and Extropic founder Beff Jezos, and said the statement warns that Anthropic is not a safe AI company. MTS is additional reporting in this story, not a second lead. The statement text itself is on the linked site.

Deep dive

CAIS, an official safety institute account, said a rare coalition across AI safety, open source, the e/acc camp, science, and law has signed the MisAnthropic Statement ahead of Anthropic's IPO, and linked a page to read or sign it. That is first hand from CAIS as the poster of the statement. The post does not quote the statement's arguments, does not give a signer count, and does not name the people in the coalition.

MTS later posted a photo and said Hugging Face cofounder Thomas Wolf, xAI cofounder Igor Babuschkin, MIRI President Nate Soares, and Extropic founder Beff Jezos have signed a statement warning that Anthropic is not a safe AI company. Those four names and that one line summary are from MTS, not from the CAIS post. They are included here as additional reporting. This file does not treat the MTS wording as the full official text of the statement.

The unusual part, if the coalition is as broad as CAIS says, is that people who often disagree about open weights and acceleration are being asked to sign the same document about Anthropic at the moment of its public listing. That is a reputational and safety fight, not a regulator's action and not a new model release.

Readers should use the statement site for the actual claims and the signer list. Until Anthropic or the exchange responds, the item is the publication of the statement, not a finding that the charges have been independently proved.

What to expect next

The statement site should be checked for the full text, the complete signer list, and any later additions. Watch for a response from Anthropic, from IPO underwriters, or from other labs named in the document. If a first hand official rebuttal or a longer CAIS thread appears, it can be added here. The statement should not be collapsed into the separate Nov 18 stories about IPO orders or cofounder wealth.