From OpenAI Engineer to Trillion-Dollar Founder: Dario Amodei and Anthropic's Five Years

Twelve years ago, Dario Amodei was just an engineer writing code for a company building China's navigation system. Two years later, he was in Silicon Valley with a small team working on a project that would go on to change OpenAI's fortunes and bring ChatGPT to the world.

But he didn't stop there. With that same small team, he founded a company in 2020 that reached a $1 trillion valuation in just five years -- a fraction of the 42 years it took Apple, the 21 years for Google, or the 17 years for Meta to hit the same milestone.

The AI model Amodei built alarmed everyone from major US corporations to the federal government, which placed it in the same export-control category as missiles and classified weapons. Ironically, reports suggest a Chinese giant like Alibaba trained its own AI on Anthropic's models without permission.

Rather than relying on the traditional approach of humans reviewing thousands of model outputs to teach right from wrong -- expensive and opaque -- Anthropic built a system where the model checks its own responses against a written 'constitution' and self-corrects. That document draws on landmark human texts, including the 1948 Universal Declaration of Human Rights, distilled into three core principles: helpful, honest, and harmless.

Anthropic is currently suing the US government, even as reports suggest parts of that same government quietly continue to use Claude -- a striking contradiction. The models Amodei built rank among the world's most powerful, but if they fell into the hands of malicious hackers, the consequences could touch everything computerized in modern life: nuclear missile controls, trillion-dollar interbank transactions, even dam gates.

That combination of extraordinary capability and potential danger is exactly why Amodei has been nicknamed 'the second Oppenheimer.'