When it comes to AI, I’m neither a pessimist nor a starry-eyed enthusiast — which is exactly why a piece Simon Willison recommended last week stuck with me. In “AI Mania Is Eviscerating Global Decision-Making,” Nik Suresh describes how AI mania is hollowing out decision-making at large companies. It’s full of sharp anecdotes from anonymous sources. And, unfortunately, very believable.
Two scenes
One example: an executive confessed he had never used ChatGPT or any AI tool in his life — right after presenting a technical strategy, entirely centered on AI, for an organization with more than $2 billion in revenue.
Even more bitter is the engineer at a company with a “token leaderboard”: he checks out a copy of the Go repository and has the AI rewrite the whole thing in Zig while he works on something else — “just so I can keep my job.”
The real question
The most interesting part is Suresh’s explanation of why the nonsense gets passed along unopposed. It’s not just sales fluff. Executives at the customers make absurd claims of 100x productivity gains. And if an executive at the vendor says those numbers aren’t plausible, it undermines the customer executive’s credibility — it’s seen as an attack, almost heresy, and it can cost the contract. So everyone stays quiet.
My take
This matters to me: it’s not an indictment of AI. I see for myself how much it delivers — my own experience plus Claude Code is a strong combination. The problem in the piece isn’t the technology, it’s hype-driven decision-making. When strategies rest on wishful numbers instead of real experience, the best model in the world won’t save you.
That’s exactly where pieces like this earn their keep. They’re a counterweight — not against AI, but against the illusion that a tool can replace your own judgment. If you actually want to use AI, you have to pick it up yourself, try it, understand it. Anything else is just a more expensive PowerPoint. And honestly? A little more courage to push back would do a lot of meetings some good.
Sources: