The CEOs of Anthropic and Google DeepMind sat down with President Donald Trump and other G7 leaders to discuss AI standards. They explained how large language models work. Trump nodded. The other heads of state nodded. No one understood a f*cking thing.
Dario Amodei and Demis Hassabis want a U.S.-led AI coalition. They pitched this idea to people who still use Yahoo Mail. The irony escaped no one except the people in the room. These are the same leaders who needed staffers to explain what TikTok does in 2023. Now they're setting policy on neural architecture.
The meeting covered AI standards. Standards for what, exactly? Standards for systems that hallucinate with confidence? Standards for chatbots that write mediocre code? The G7 will standardize the thing that changes every six months and that none of them can define without reading from a script.
Retail traders heard "AI coalition" and immediately bought call options on semiconductor ETFs they cannot spell. They think this means NVIDIA goes up. It does not mean that. It means seven countries will form a committee that forms a subcommittee that publishes a PDF no one reads.
Trump's presence adds a layer of absurdity that feels almost too easy. The man who asked if we could nuke a hurricane now helps set guardrails for artificial general intelligence. The simulation is getting lazy with its writing.
Amodei and Hassabis walked out of that meeting knowing they accomplished nothing. They know the real decisions get made in Palo Alto and London, not at a photo op in Italy. But they smiled for the cameras anyway. They said the right things about cooperation and safety.
The coalition will produce zero enforceable policy and several strongly worded statements about the importance of democratic values in AI development, which China will ignore while building better models.
Photo by on Unsplash
Leave a Comment