Dario Amodei, 44, is the CEO of Anthropic. In recent days he has launched an appeal to slow down the development of the most advanced artificial intelligence models: “AI Agents could take control of the internet within six months”. The problem is that this choice could favor companies that already lead the market.
Anthropic and OpenAI are the two private companies with the highest estimated value in the world. With this formula we define the companies that are not listed on the stock exchange. It is a condition in which SpaceX also found itself, at least until June: now its market capitalization has reached 1,993 billion dollarsthe seventh highest in the world. For Anthropic and OpenAI the data fluctuates and changes based on who publishes the rankings. We can take a chance a solid average: Anthropic is rated 960 billion dollarsOpenAI stops at 890 billion dollars. Let’s start from here, to read everything that has been said in the last few days. The news is known. Dario Amodei, CEO of Anthropic, has launched an appeal to slow down the development of models based on artificial intelligence. A position clarified by Amodei himself in an interview with the US television network CBS: “We must ensure that the growth of AI follows our ability to control, otherwise AI Agents could take control of the internet within six months.” In support of this thesis there is a report published by Anthropic itself on the hacker groups that used Claude’s models to plan cyber attacks.
The appeal was also shared by other names that weigh heavily in the sector. First of all Sam Altmanfounder of OpenAI but also Demis Hassabispresident of Google DeepMind and Nobel Prize winner. Last but not least, Elon Musk also signed the appeal. His Grok models still fail to compete with others but the cards to participate in the race are there. Amodei’s proposal it is precise: among other things suggests that the laboratories where artificial intelligence is developed are checked by independent auditorswhich can certify standards and risks of the algorithms studied. An external regulatory body, such as theInternational Atomic Energy Agency which deals with the production of nuclear weapons.
Noble intent but it’s hard not to see the implication. At this moment the main models linked to artificial intelligence all belong to the same hub: the companies mentioned are based in the United States and there are links between them that have developed over the years. Dario Amodei previously worked at OpenAI. Part of the funding Sam Altman used to develop ChatGPT came from Elon Musk. Trump does not agree, for him stopping means leaving room for competition. Other entrepreneurs, however, are explaining these days that they should leave establish rules and access criteria for those who today lead the development ofArtificial intelligence risks creating a cartel. Control mechanisms such as those proposed by Amodei risk making market access more difficult for small companies.
Now the AI has become a bomb to be disarmed: Amodei’s move to buy time and increase controls
If he asks Aidan GomezCEO of the Canadian startup Cohere. As reported by the New York Times, Gomez published an essay attacking this oligopoly: “Is it right that a handful of market-leading Silicon Valley artificial intelligence companies can define the rules and security standards of a revolutionary technology for the entire world? These oligopolies now ask to circumvent the rules of competition and to be able to dictate conditions to everyone else.”
