From "anti-human" to AI leader, Musk completely changes his tune on Anthropic

According to Dongcha Beating monitoring, Musk has publicly admitted that he previously misjudged Anthropic. It is now the leading force in the AI field, and no company has released better models than Mythos and Fable.

This turnaround is no small shift.

In February this year, Musk also said that Anthropic is “hostile to Western civilization.” Its models are hostile toward white people, Asians, heterosexuals, and men, and it is an “anti-human and evil” company. He also accused Anthropic of massively stealing training data, describing the company as “self-righteous, sanctimonious, and hypocritical.” In March, he again asked: “Is there a more hypocritical company than Anthropic?”

The change came after the two sides began doing business.

In May, Anthropic leased SpaceX’s Colossus 1 data center, which can support more than 220,000 Nvidia GPUs. Musk then immediately said that he spent a great deal of time communicating with Anthropic’s top leadership and found that they were “very capable and also very concerned about doing the right thing,” and that no one had triggered his “evil detector.”

Two months later, his “evil detector” had already been upgraded into a championship certification. Musk not only ranked Anthropic as number one in AI, but also promised not to suddenly cut off computing power to harm them, because “that’s not my style.”