Anthropic CEO Dario Amodei’s volatile personality and penchant for public rants has sparked growing concerns among the AI giant’s shareholders – some of whom fear he’s ill-equipped to steer the $380 billion company through its current troubles, The Post has learned.
Amodei, whose firm built the “Claude” chatbot, turned heads last month when he blasted President Trump and OpenAI’s Sam Altman in an explosive internal memo. It was sent hours after the Pentagon’s decision to blacklist Anthropic for refusing to remove safeguards on how its AI can be used in military settings.
In the memo — which was promptly leaked to media — Amodei claimed the Pentagon targeted Anthropic for not giving “dictator-style praise to Trump (while Sam has).” He likewise accused Altman of telling “straight up lies” by claiming to share Anthropic’s safety concerns, and called OpenAI employees a “gullible bunch.”
One current Anthropic shareholder, who spoke to The Post on condition of anonymity, said Amodei’s fiery diatribe was part of an “extremely concerning” pattern that’s unbefitting of a high-profile CEO whose firm counts Amazon and Google among its blue-chip investors.
“You are a f—king CEO who has raised billions of dollars. You can’t just rant and expect all shareholders to have the same mentality that you have,” the shareholder said.
“The one thing that was striking to me is the aggressiveness of his opinions on Trump and the Pentagon – given that they were the only ones whose tech was being used in things at the Pentagon,” the shareholder added.
The brilliant-but-eccentric CEO has often courted controversy in his public appearances – once warning that AI could drive unemployment to 20% and likening the White House’s decision to allow sale of advanced AI chips to China to “selling nuclear weapons to North Korea.”
Amodei has a habit of speaking out despite internal efforts – usually by his sister and fellow cofounder Daniela and policy chief Jack Clark – to restrain his impulses, the shareholder added.
“They try to reel him in but he clearly cannot control his emotions,” the shareholder said.
A source close to Anthropic pushed back on the criticism of Amodei, noting that the company remains one of the fastest-growing in history and has received public support from several investors, including Altimeter, Menlo Ventures and Spark Capital, following its dispute with the Pentagon.
Amodei’s particular disdain for Altman has simmered since before he and a handful of close allies left OpenAI to found Anthropic in 2021.
The feud was evident during a painfully awkward incident at a February tech conference in India. The two rivals were placed next to each other on stage alongside Indian Prime Minister Narendra Modi and Google CEO Sundar Pichai, but refused to link hands for a group photo as the crowd cheered.
Dario and Daniela Amodei exited OpenAI after having disagreements with Altman, who they felt was recklessly prioritizing business growth over AI safety guardrails. Those concerns went into overdrive after Altman decided to release ChatGPT to the public in 2022.
Last week, the Wall Street Journal reported that the Amodei siblings had workplace tensions with Altman and fellow OpenAI cofounder Greg Brockman. One blowup occurred in late 2020, when Altman accused the siblings of urging colleagues to badmouth him to OpenAI’s board.
Behind the scenes, some insiders, who spoke to The Post on condition of anonymity, say Amodei remains highly critical of OpenAI and obsessed with beating Altman – to the point that some researchers fear it’s hurting the company’s actual mission.
The Anthropic boss is also openly disdainful of OpenAI employees, saying they are “under Sam’s spell” and untrustworthy as a result, the shareholder said.
“It is very clear it is a personal vendetta, but their justification is their core conviction that Sam can’t be trusted with prioritizing trust and safety,” the shareholder added.
One well-connected industry source described Amodei’s feelings toward his former boss-turned-rival as “Sama Derangement Syndrome” – a reference to Altman’s account handle on X, @Sama.
Anthropic declined to comment.
Despite the upheaval, Amodei maintains a loyal following inside Anthropic. The 43-year-old is known to host regular meetings that are referred to internally as “Dario Vision Quests,” or DVQs, where he holds forth on everything from AI policy and company business to geopolitics and beyond.
Amodei has since apologized for the Pentagon-bashing memo, stating that “Anthropic did not leak this post nor direct anyone else to do so” and that it was “not in our interest to escalate this situation.”
“Any normal CEO would know that like, f—k, I’m risking a lot by saying this all in writing,” the shareholder said. “But he’s blinded by his own love of self and intelligence and those around him who are like ‘yes, you are God.”
Anthropic is currently suing the Pentagon over the “supply chain risk” designation and described the US government’s actions as “unprecedented and unlawful.” A top Pentagon official argued the tag was necessary because the Claude AI chatbot was trained using a fundamentally different ideology from what the Pentagon wants for its systems.
A federal judge issued a preliminary injunction blocking the government ban on March 26.
The source close to Anthropic said the lawsuit was necessary to protect the firm’s business interests, but it is still focused on working alongside with the government.
On the surface, Anthropic’s pursuit of the Pentagon contract in the first place seems to contradict its AI safety obsession.
However, sources familiar with the firm’s thinking say it was part of Amodei’s attempt at ensure his firm was at the forefront of the debate on how and when advanced AI is deployed on the battlefield.
Critics, including former White House AI czar David Sacks, have long alleged that Anthropic’s safety branding is actually an elaborate attempt at “regulatory capture” – Silicon Valley lingo for crafting the rules in such a way that they benefit and their rivals struggle.
Reuters previously reported that investors had expressed support for Amodei during the Pentagon spat, even as some privately expressed frustrations that he had antagonized the Trump administration with his rhetoric as the situation spiraled.
On the eve of a Pentagon deadline to strike a deal, Amodei publicly announced that Anthropic could not “in good conscience accede to their request.”
“It’s an ego and diplomacy problem,” one source told the outlet.
Credit: Source link
