On Tuesday the heads of the big AI companies stood around Donald Trump outside the White House and produced the image of an industry tackling the public’s fears together. The Wall Street Journal now reports what happened just before, behind the doors: industry leaders, Nvidia’s Jensen Huang among them, had questioned Dario Amodei about his warnings over the dangers of the technology — according to people familiar with the matter.
The piece, by Amrith Ramkumar and Josh Dawsey, is marked exclusive and sits behind the paywall. Beyond those two public paragraphs there is nothing here I was able to read myself.
Why this is more than an anecdote
The sequence is the point. The agreement to sign the statement of principles was already in place. The questions to Amodei came after that. And only then the group photo.
Huang and Amodei disagreeing is not news. Huang has put the odds of AI-driven doomsday at zero percent; Amodei has spent years warning in both directions — about the upside that gets lost and the damage that can follow. What’s new is that the argument has moved out of interviews and into the room where the voluntary commitment was negotiated.
The unity is the photo, not the paper
Four layers of oversight and an AI officer sit in an accord that compels nobody to do anything. What holds it up is the assumption that the signatories want the same thing. The Journal just put a scratch in exactly that — and not through an outside critic, but through the people standing in the picture.
For Anthropic this is the most awkward place to be: signed up, and explaining itself to its own peers. Amodei’s week started with dinner at Trump’s table and ends with an FTC investigation. In between he stood in the White House and had to justify why he puts it so bluntly.
Sources: