Anthropic's CEO answered a safety critic with a 2,000-word post on X, arguing the company does more good by staying at the front of the AI race.
Dario Amodei, the CEO of Anthropic, the company behind the Claude AI assistant, wrote a nearly 2,000-word reply on X to a critic who said the company was putting safety in the back seat. His answer: Anthropic does more for safety by staying at the cutting edge than by stepping aside.
On Monday, Nathan Calvin posted a list of nine things Anthropic had done in the past year that he said undermined its safety promises. Calvin is chief strategy officer of Encode AI, a nonprofit that advocates for AI safety. His post ran over 1,500 words. He said Anthropic was trying to win a race to superintelligence, meaning AI far smarter than people, and was giving safety less priority.
On Tuesday, Amodei said he was a fan of Calvin's investigative work but disagreed with many of his points. He called Anthropic's approach a "race to the top." The idea is that by showing it is possible to build the technology responsibly, the company can raise the bar for the whole industry.
He then posted his own list of nine actions that he said hurt Anthropic commercially but had strong safety benefits. Two examples:
Amodei did concede that Anthropic could have handled some issues better. One was a draft bill the company circulated to Congress this month, first obtained by Politico. Critics said it would make it harder for the federal government to block dangerous AI models. He still wrote that he is "extremely proud" of the company's track record and stands by it.
The debate comes as safety worries grow. In September, AI researchers predicted an "AI apocalypse," in which increasingly capable and rogue technology could kill all of humanity within a decade. Amodei, OpenAI CEO Sam Altman and SpaceXAI CEO Elon Musk all agreed to "pace the frontier," which means slowing development so enough checks are in place.
Money is also in play. Anthropic is on the brink of an IPO, when a company first sells shares to the public. It is expected to list in November at a $2 trillion valuation. Last week it reported third-quarter revenue of $16.8 billion.
The material does not say whether Calvin replied, or whether the draft bill will change. Whether Anthropic's approach actually makes the industry safer is the open question, and the two sides still disagree on it.
Anthropic says competing at the frontier makes AI safer, critics say it does the opposite, and the argument is heating up before its expected November IPO.
OpenAI has about $11 billion in hand at a $1.25 trillion value, below its first ask. Microsoft and NVIDIA shareholders have reason to watch what comes next.
Read more: OpenAI raises $11 billion, but at a lower price than it wantedVisa tested 22 AI models on realistic shopping tasks. They passed just over half, and often bought the wrong thing anyway.
Read more: Visa tested 22 AI shoppers, and most still make costly mistakesDARPA wants AI to hunt for design flaws in stock, bond, energy and cloud markets, so enemies cannot use them to cause economic damage that reaches households.
Read more: Pentagon wants AI to find weak spots in stock and energy markets
