Futurology

I lot of the critics were right about it for all the wrong reasons. You don't get to claim victory on that, it's actually better to make a good argument that later turns out to have a fatal flaw.

I don't know this exact guy's history, though. He might not be in the picture. Here's the article he mentioned from 3 years ago. He makes some pretty good points, and predicts something like chain-of-thought, although I have a feeling radiology AI has come a long way. The post shown about the market trajectory from about a year ago was also spot on.

Tangentially, I fully believe a human whitehouse staffer chose to use top-level internet domains to break the tariffs down. An LLM would done something more bespoke. If you have shitty resources and an unclear request it's not even a terrible approach.