Unelected Men are deciding on AI

Anthropic's launch of its most powerful public model, Fable 5, lasted only days before the US Commerce Department issued an unprecedented export ban. Anthropic deactivated the model entirely and pushed back, arguing the flagged capabilities already exist in competing models.

The backstory is messy. Amazon reportedly helped trigger the intervention — possibly to win favor with Trump for defense contracts. Defense Secretary Hegseth has called Anthropic's CEO "an ideological maniac," while OpenAI's president is one of Trump's biggest donors. As I told De Volkskrant: US AI policy looks like it's being driven by unelected tech bros calling the president to undermine their competitors. You reject Anthropic but embrace Palantir? None of it is coherent.

Anthropic plays a role too. When you position yourself as the liberal good guys and market your product as a cyber nuclear weapon, you shouldn't be surprised when it provokes exactly this kind of response.

For Europe, the lesson isn't to blindly join the AI arms race. Should we also spend €1.6 trillion on data centers? China's approach — deploying AI strategically in specific industries rather than chasing superintelligence — is more attractive.

Read the whole article here