AI mania is eviscerating global decision-making

Nik Suresh, writing as Ludicity, has a knack for naming the thing everyone in the meeting is thinking and no one will say out loud. His latest — which shot up Hacker News over the weekend (via Simon Willison) — is about how “AI strategy” actually gets made inside large organizations: not from experience or evidence, but from hype. The anecdote that lands hardest is an executive who had never once opened ChatGPT authoring an AI-centric technical strategy for a company north of $2B in revenue.

His sharpest point is about incentives, not models. Leaders can’t say what they actually believe about how to steer their organizations without putting their jobs at risk, so everyone performs enthusiasm instead — a “distributed government by assassination,” as he puts it.

I read a lot of AI writing that is either breathless or reflexively sour, and this is neither. It’s a working consultant describing, with specificity and a lot of jokes, the gap between what gets decided and what anyone has actually tested. That gap is the whole game in my line of work too — a claim isn’t true because the deck says so — and it’s rare to see it drawn this cleanly.