Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-07-19 · high
Nik Suresh says inflated AI productivity claims survive because vendor executives can't contradict clients without risking their contracts.
Why now: Willison linked Suresh's essay on July 19, pushing an incentive-structure argument past the usual AI hype cycle takes.
An executive wrote a $2 billion company's AI strategy without ever using an AI tool, according to an essay by Nik Suresh that Simon Willison linked on July 19.
Suresh's sharper point isn't about that one executive. It's about how AI productivity numbers get inflated in the first place. Customer-side executives report 100x gains. Vendor executives who know better can't contradict them, because doing so torches contract credibility and, often, their own jobs. Nobody in that chain is lying exactly. Nobody has an incentive to correct the number either. The inflation just perpetuates itself through people protecting their positions.
The detail that lands hardest: engineers rewriting working codebases in different languages, not because the rewrite serves the product, but "to keep my job." That's a narrower failure than an executive skipping the tools, and a more useful one. It's what AI adoption pressure looks like when people get rewarded for appearing productive instead of being productive.
Every 100x productivity claim you hear secondhand is probably a number nobody checked, because nobody in the reporting chain gets rewarded for checking it. Vendors profit from the inflated figure holding. Customers who reported it don't want to walk it back. Engineers under pressure to show AI usage will optimize for looking busy with it over shipping with it. Watch whether the gap between reported and audited AI gains widens before anyone with the incentive to close it shows up.
Willison linked the essay on July 19, giving an incentive-structure argument more reach than a typical hot take about AI hype.
Each link below shares sources, entities, or timing with this story.
Simon Willison released Datasette / Shared entities / Same source domain / Earlier coverage / Tension
Linked by a graph relationship (Simon Willison released Datasette); both cover Simon Willison, Willison; reported by the same outlet (simonwillison.net).
Simon Willison uses Claude / Shared entities / Same source domain
Linked by a graph relationship (Simon Willison uses Claude); both cover July, Simon Willison, Willison; reported by the same outlet (simonwillison.net).
Simon Willison released LLM / Shared entities / Same source domain / Earlier coverage
Linked by a graph relationship (Simon Willison released LLM); both cover Simon Willison, Willison; reported by the same outlet (simonwillison.net).
Linked by a graph relationship (Simon Willison released LLM); both cover Simon Willison, Willison; reported by the same outlet (simonwillison.net).
Simon Willison released Agentic Engineering Patterns / Shared entities / Same source domain / Earlier coverage
Linked by a graph relationship (Simon Willison released Agentic Engineering Patterns); both cover Simon Willison, Willison; reported by the same outlet (simonwillison.net).
Simon Willison uses Claude / Shared entities / Same source domain / Earlier coverage
Linked by a graph relationship (Simon Willison uses Claude); both cover Simon Willison, Willison; reported by the same outlet (simonwillison.net).
Simon Willison uses Claude / Shared entity: Simon Willison / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Simon Willison uses Claude); both cover Simon Willison; reported by the same outlet (simonwillison.net).
Linked by a graph relationship (Simon Willison uses Claude); both cover Simon Willison; reported by the same outlet (simonwillison.net).