13
Five men control AI. Who should control them? | The Economist Insider
www.economist.comIt’s the age of the AI demi-god. Dario, Demis, Elon, Mark and Sam—five tech bosses so iconic you need only use their first names—sit atop the world’s leading AI firms, building the technology that will define the future. America’s government has left them to it, wary of anything that might slow its tech race with China. Now, that looks set to change. The startling capabilities of Anthropic’s new model, Mythos, and a wave of public anxiety about AI have alarmed the Trump administration. It now must wrestle with a profound question: how can you govern a transformative technology without killing it?
Zanny Minton Beddoes, editor-in-chief, Edward Carr, deputy editor, and a panel of our expert journalists weigh the dangers of unchecked AI power against the risks of heavy-handed regulation—and ask who should ultimately be in control.


I’m pretty surprised to see Zuckerberg included in this group in 2026. His company is totally irrelevant in the AI space now since the only relevancy they had was due to releasing their model weights publicly.
The llama 4 model was a complete failure that was already obsolete on release, and since then, they’ve stopped releasing model weights to the public. They haven’t shown any ability to produce a competitive model since llama 3, which is now very old in AI timescales.
It’s really interesting that they blundered so hard, because releasing model weights was a huge win for them that gave them early dominance over local and open source AI tooling and workflows. Most of their talent has abandoned ship, and now they have no talent, no competitive models, and no marketshare of local AI, right at a time when closed/cloud AI providers are becoming prohibitively expensive and inaccessible and local AI is taking off.
They’ll probably be a textbook case study in how to fuck up, someday.