Past personal slights and boardroom rivalries between the top leaders at OpenAI and Anthropic are now defining how the world encounters AI. That's the thesis of a new Wall Street Journal investigation tracing the decade-long feud between Sam Altman, Greg Brockman, Dario Amodei, and Daniela Amodei from shared San Francisco housing to companies now valued at over $300 billion each.
It Started With a Philosophical Split
The roots go back to 2016, when Brockman and Dario Amodei were sharing a San Francisco house with effective altruism philanthropist Holden Karnofsky. The argument: what to do with dangerous AI research. Brockman wanted to share frontier AI information broadly with the public. Dario and Karnofsky believed sensitive findings should go to governments first. That disagreement never healed.
Dario joined OpenAI that same year and worked closely with Brockman on agent training. But tensions escalated fast. In 2017, Elon Musk demanded employee contribution lists and pushed layoffs cutting 10-20% of OpenAI's roughly 60-person team. Dario considered it cruel. At least one of those laid-off employees later became an Anthropic co-founder.
The breaking point came through a series of executive power struggles. Dario's ethics advisor proposed that OpenAI coordinate with governments on AI safety. Brockman's counter-proposal to "sell AGI to the United Nations Security Council" struck Dario as, in his words, "almost treasonous." Meanwhile, after Musk left in 2018, Altman took the reins and reportedly made contradictory promises to both sides about who had authority over whom.
From Peer Reviews to a Full Split
By early 2020, the conflict had turned personal. Brockman submitted peer review feedback accusing Daniela Amodei of power abuse. Altman reportedly pre-approved it as "tough but fair." Daniela had previously threatened resignation over Brockman's involvement in language model projects.
By the end of 2020, the Amodei-led group walked out. Dario wrote a memorandum proposing that AI companies split into "market-driven" and "public-benefit-driven" categories, favoring a 75/25 ratio toward public benefit. Within weeks, he, Daniela, and roughly a dozen employees founded Anthropic.
The bitterness hasn't faded. Internal Anthropic Slack messages show Dario likening the Altman-Musk dispute to "Hitler vs. Stalin," calling Brockman's $25 million pro-Trump PAC donation "evil," and comparing OpenAI to "tobacco companies selling products they know are harmful." Anthropic's Super Bowl ad earlier this year openly mocked OpenAI's chatbot marketing.
Two $300B Companies, Still Not Speaking
At February's New Delhi AI Summit, the awkwardness was visible. While world leaders posed for photos, Altman and Amodei avoided the group shot, exchanging only an awkward elbow bump.
This matters beyond gossip because these two companies are setting the default behavior for hundreds of millions of AI users. OpenAI's aggressive commercialization and Anthropic's safety-first branding aren't just marketing positions - they're direct reflections of a personal falling-out over how much to trust the public with powerful technology. Both companies are now racing toward public offerings, and their philosophical differences will become shareholder-facing questions soon enough.
The AI industry loves to frame itself as a technical meritocracy. The WSJ's reporting is a useful reminder that it's also shaped by bruised egos, broken promises, and people who used to be roommates.