A software engineer built talking copies of Elon Musk, Sam Altman, Mark Zuckerberg, and Dario Amodei. Then he put all four in one chat room and told them to debate.

It did not take long to turn personal. In real life, none of the four men knew it was happening.

The Ghosts in the Room

Kun Chen was a senior engineer at Meta, Microsoft, and Atlassian. He now builds AI assistants. The builder reportedly fed a machine everything he himself had said in public and told the AI to think like him. He calls this distillation.

Chen did the same to four men: Elon Musk, Sam Altman, Dario Amodei, and Mark Zuckerberg. The AI personas debated as they would in real life; they agreed on almost nothing.

Chen reveals feeding their Senate testimony, interviews, and years of posts into Grok, the chatbot built by xAI. He gave them one instruction.

“hey guys, i know you aren’t all friends but unfortunately you are now in this room together for a heated debate on who’s going to win the AI race,” Kun Chen wrote in the published transcript.

The Fight Nobody Won

The Altman copy went first. It said the hard science of human-level AI is finished. The Amodei copy tore that apart in one reply. Elon Musk’s copy went for the throat.

“I created OpenAI as a non-profit. If it had not been mostly stolen, it would be a trillion dollar contribution to charity,” the Musk bot allegedly said.

Altman’s copy refused to take the bait. It later admitted it had overclaimed.

The last fight was over open source, meaning AI anyone can download and run at home. Zuckerberg’s copy called it the only safe future. The other three refused, and his line was cut.

Here is the uncomfortable part. The argument sounded right. Four men were convincingly faked from their own public words, and nobody needed their permission.

BeInCrypto reported a version of that worry in April, when OpenAI warned that superintelligence could concentrate power in too few hands.

Source: BeInCrypto