Дарио Амодеи: AI-подросток вот-вот разнесет наш уютный мир

Dario Amodei seems to have decided that short social media posts are not enough to prepare us for the coming changes. When the head of one of the world's most expensive and influential AI laboratories releases a text the size of a short novel, that's reason enough to at least set aside your work and read carefully. The Anthropic co-founder published an essay titled "The Teenage Years of Technology," and it's not just leisure musings, but a full-fledged manifesto from a man who sees the code of the future before anyone else.

Amodei is no newcomer to security issues. He left OpenAI precisely because he considered Sam Altman's approach too reckless in terms of risks. Now, at the helm of Anthropic, he openly declares: humanity is entering a phase that will test us as a biological species. His metaphor of the "teenage years" of technology perfectly describes the current chaos. We have already given artificial intelligence enormous power, but we have not yet taught it (or ourselves) responsibility. This is the very moment when a teenager can already get behind the wheel of a sports car, but hasn't yet realized that the concrete wall ahead is not a video game prop.

The main question that Amodei poses: are our social and state institutions ready for the "almost unimaginable power" that will become reality in the very near future? He speaks of this not as a theoretical possibility fifty years from now, but as an event that is literally knocking on the door. While the public debates whether neural networks will replace copywriters or designers, Amodei is looking at fundamental shifts in biology, cybersecurity, and government. He sees AI as a tool capable both of accelerating scientific progress by centuries and of creating risks to which we have not developed immunity.

The irony of the situation is that Anthropic itself creates some of the most powerful models on the market. Claude 3.5 Sonnet has already caused many developers to question GPT-4's leadership. A strange paradox emerges: a man builds the fastest machine in the world and simultaneously writes 19,000 words about why fast driving can kill us. It's not hypocrisy, but rather an attempt to sit on two chairs at once—to remain a commercially successful player while preserving the status of "industry conscience."

British polls meanwhile show that a quarter of the country's population seriously fears losing their jobs due to AI in the next five years. But for Amodei, this is just the tip of the iceberg. What concerns him is not simply the labor market, but the existential sustainability of human systems. Will we be able to control intelligence that surpasses our own? Will our democratic procedures turn into slow and useless ballast in a world where decisions are made in milliseconds? Amodei doesn't provide simple answers, but his tone leaves no doubt: the time for quiet contemplation is over.

This manifesto looks like an attempt to set the agenda for regulators. Amodei understands that if the industry doesn't establish the rules of the game now, the consequences will be uncontrollable. He calls on the world to "wake up," hinting that we are still in a state of pleasant half-sleep, enjoying how neural networks write poems for us and draw pictures. But the real power he warns about has nothing to do with entertainment. It's a question of the redistribution of power on a global scale.

Bottom line: Anthropic finally solidifies its role as the "adult in the room" who isn't afraid to spoil the party with scary stories. The question is only whether the other Silicon Valley giants will hear this call, or whether the AI arms race will ultimately move into a stage of uncontrollable dive.