AI can do more of the work.

It cannot decide what is worth doing.

This week’s That Was The Week asks the civilization question behind the AI boom: if machines take over more execution, what do humans choose to build with the time, capital, and freedom that remain?

Three narrative engines are competing to define what AI means: the doom industry (selling fear), the corporate machine (selling story), and the financial machine (selling access). All three assume ordinary people need mediators to understand AI. All are driven by financial or political incentives. But the actual evidence from this week — chatbots that moderate rather than radicalize, a smartphone panic that collapses on cross-cultural data, real security risks that are concrete and addressable — suggests humans are more capable than any of these narratives give them credit for. The question isn't whether AI needs to be explained to people. It's whether anyone will let them think and act for themselves.

Growing Up? Winning Wars Involves Losing Battles.

Anthropic won a First Amendment ruling against the Pentagon. OpenAI killed Sora. One insisted on principle. The other chose discipline.

Meanwhile: software trades below the S&P 500 for the first time ever, Jensen pitched a trillion-dollar token factory, and David Sacks left the building.

Intelligence is getting cheaper. The question is who earns trust while it happens.

This week's That Was The Week: https://thatwastheweek.com