Sam Altman Critiques Anthropic Amid Feud with Elon Musk

Sam Altman Critiques Anthropic Amidst Feud with Elon Musk

Sam Altman, the CEO of OpenAI, recently voiced sharp criticisms against Anthropic, a competing company in the field of artificial intelligence. This public attack comes at a time when Altman is already embroiled in legal conflicts with Apple and tensions with Elon Musk. The feud with Anthropic was sparked by a promotional video the company released, which is part of its new social initiative. Altman accused Anthropic of quietly restricting access to its AI tools, escalating tensions between the two companies.

A Controversial Video from Anthropic

The video in question was published on the X account of Claude, a platform from Anthropic. It featured a montage of various images accompanied by voiceovers posing deep questions about the future of AI, such as "Can we trust AI?" and "If it ends up taking almost all jobs, what does it mean to work?" Titled "There is Hope in Difficult Questions," this video is part of Anthropic's "Inviting Difficult Questions" initiative. The company's stated goal is to address major public concerns regarding AI while promising to be transparent about the actions taken to respond to these questions.

Sam Altman's Cutting Response

In response to this initiative, Sam Altman expressed his skepticism. In a message posted on X, he stated that he initially took the video for satire, even searching for a humorous pseudonym like "c1audeai." Altman criticized Anthropic's approach, asserting that asking difficult questions only makes sense if one is deemed worthy of accessing the answers without facing restrictions. This statement echoes the criticisms that Anthropic has received regarding the limitations imposed on its AI model, Fable 5.

Anthropic's Safeguards Under Fire

Anthropic has indeed implemented safeguards around Fable 5, its advanced AI model. These measures aimed to redirect sensitive questions on topics such as cybersecurity, biology, and chemistry to less capable models. The goal was to prevent distillation attacks, where weaker models are trained on the outputs of advanced models. However, these restrictions have faced criticism, particularly because they were not initially communicated to users. In response to pressure from AI researchers, Anthropic revised its policy on June 10, now informing users when their requests are limited.

A Context of Multiple Tensions for Altman

Altman's critique of Anthropic comes in a context where the OpenAI CEO is already under pressure. Apple recently filed a lawsuit against OpenAI, accusing it of poaching its former employees and accessing confidential information. Simultaneously, Altman is engaged in a public dispute with Elon Musk. The latter referred to Altman as "Scam Altman" in reaction to the Apple lawsuit, to which Altman responded by accusing Musk of deception with his plans for space data centers.

A Persistent Rivalry Between OpenAI and Anthropic

The rivalry between OpenAI and Anthropic, as well as between their respective leaders, Sam Altman and Dario Amodei, is not new. Since Amodei left OpenAI in 2020 to found Anthropic, tensions have continued to rise. Incidents such as their refusal to shake hands at a summit in India or the simultaneous announcement of their intentions to take their companies public illustrate this rivalry. There is little indication that this competition will ease anytime soon.