A private feud lasting 10 years, if not for OpenAI's "hypocrisy," would not have led to the world's strongest AI company, Anthropic

Original Title: 《The Decadelong Feud Shaping the Future of AI》

Original Author: Keach Hagey

Wall Street Journal reporter Keach Hagey published a lengthy investigative report, revealing for the first time the decade-long personal feud between the founders of Anthropic and OpenAI through extensive interviews with current and former employees and executives of both companies. What shapes the global AI landscape is not just a technological rivalry but also an unhealed personal trauma.

In recent months, Dario Amodei's internal rhetoric has been much more intense than in public. He compared the legal dispute between Sam Altman and Elon Musk to "the struggle between Hitler and Stalin," called OpenAI president Greg Brockman's $25 million donation to a pro-Trump super PAC "evil," and likened OpenAI and other competitors to "tobacco companies that knowingly sell harmful products."

After the Pentagon dispute escalated, he referred to OpenAI as "hypocritical" on Slack, writing, "These facts indicate a behavioral pattern I frequently observe in Sam Altman."

Internally, Anthropic refers to this branding strategy as creating a "healthy alternative" to competitors. During this year's Super Bowl, an unnamed advertisement satirizing OpenAI for embedding ads in chatbots was a public product of this strategy.

The story begins in the living room of a shared house on Delano Street in San Francisco in 2016. Dario and his sister Daniela Amodei lived there, and OpenAI co-founder Brockman frequently visited due to his close relationship with Daniela. One day, Brockman, Dario, and Daniela's then-fiancé, effective altruist philanthropist Holden Karnofsky, sat together debating the correct development path for AI: Brockman believed all Americans should be informed about what was happening at the forefront of AI, while Dario and Karnofsky argued that sensitive information should be reported to the government first rather than broadcast to the public. This disagreement later became a watershed moment in the philosophical direction of the two companies.

Moved by OpenAI's talent pool, Dario joined in mid-2016, staying up late with Brockman to train AI agents to play video games. However, after four years of working together, tensions surrounding power and a sense of belonging deepened. In 2017, when major OpenAI investor Musk demanded a list of each employee's contributions for layoffs, about 10% to 20% of the 60-person team was let go one by one, which Dario viewed as cruel; one of those laid off later became a co-founder of Anthropic.

That same year, an ethics advisor hired by Dario suggested that OpenAI act as a coordinating entity between AI companies and the government, from which Brockman extrapolated the idea of "selling AGI to the nuclear powers of the UN Security Council." Dario considered this nearly treasonous and contemplated resigning.

After Musk stepped down in 2018, Altman took over leadership. He and Dario reached a consensus: employees lacked confidence in the leadership of Brockman and chief scientist Ilya Sutskever. Dario agreed to stay on the condition that the two would no longer be in charge, but soon discovered that Altman had simultaneously promised the latter two that they had the right to fire him, creating a contradiction between the two commitments.

After the development of the GPT series was initiated, the executive team erupted into the most intense conflict over who could participate in the language model project. Dario, then the research director, did not allow Brockman to interfere, while Daniela, who co-led the project with Alec Radford, threatened to resign from her position. Radford's personal wishes became entangled in the executive proxy war.

Dario's credentials rose with the success of GPT-2 and GPT-3, but he felt that Altman downplayed his contributions. When Brockman discussed OpenAI's charter on a podcast, Dario was angry for not being invited despite having contributed more to the charter; he was also dissatisfied to learn that Brockman and Altman were meeting with former President Obama while excluding him.

The conflict escalated dramatically during a confrontation in a meeting room. Altman summoned the Amodei siblings and accused them of inciting colleagues to submit negative feedback about him to the board. The two denied it. Altman claimed the information came from another executive, and Daniela immediately called that executive in for questioning, who said he was completely unaware.

Altman then denied having said that, leading to a heated argument. In early 2020, Altman asked executives to write peer reviews of each other, and Brockman wrote a harshly worded feedback accusing Daniela of abusing power and using bureaucratic processes to exclude dissenters, which Altman reviewed in advance and described as "tough but fair." Daniela rebutted each point, escalating the argument to the point where Brockman suggested retracting his comments.

By the end of 2020, a team centered around Dario decided to leave, with Daniela leading negotiations with lawyers regarding their departure. Altman personally visited Dario's home to persuade him to stay, but Dario insisted he would only accept reporting directly to the board and clearly stated he could not work with Brockman. Before leaving, he wrote a lengthy memo categorizing AI companies into "market-oriented" and "public interest-oriented," arguing that the ideal ratio was 75% public interest and 25% market. Weeks later, Dario, Daniela, and nearly a dozen employees left OpenAI to found Anthropic.

Five years later, both companies are valued at over $300 billion and are competing to go public. At the conclusion of the AI summit in New Delhi this February, Indian Prime Minister Modi and the tech leaders present raised their hands together, while Amodei and Altman chose not to participate, merely awkwardly bumping elbows.