The next chapter of my book-in-process, The Mostly Helpful Psychopath, is now live. The Chinese Room, AI consoles my canine and existential woes, and proves to be a able replacement for an encyclopedia.
A CEO told me his firm felt like a ship with barnacles on it. He was right...but you can't solve for barnacles. Here's why the models that sound best fit worst, what Huge's public experiment proved, and two exercises that help you convert the fog into a real diagnosis.
Forty years ago I caught myself talking to my dead mother on a hiking trail, narrating the plants to someone who wasn't there.
That moment turned out to
China just gave away an AI model that rivals America's best. The real danger isn't who's ahead — it's what happens when anyone can download it, reshape its worldview, and put it on a phone they sell for next to nothing.
Every time your team uses AI where the training data is thin, the output arrives fast and the tax follows: evaluating, correcting, reworking. Workday's 2026 research says 37–40% of the time AI saves gets eaten by checking its work. I think they're optimistic. Here's why.
Altman’s judgment day came and went in a single interview. Yours arrives every morning, in the thousand small calls no machine can make for you. For the leaders willing to do the thinking, it’s the real opportunity.
Your best clients can't make a decision right now, and you're misreading why. It isn't caution about budget. They're frozen, and they're waiting for the machine to hand them judgment it can't give. That's the paradox sitting under the paralysis.
Most agencies that lose a long-standing client never see it coming. The work was solid. The relationship was warm. And then the client just… didn't need them anymore.
