What's this about?
Mollick predicts polls will show that everyone hates AI overall while everyone secretly uses it constantly, and that AI companies will be underwater in public opinion while many individuals feel strong personal attachment to their favourite model.
What exactly is the contradiction?
It is the gap between stated and revealed preference. People answer surveys about a category — "AI" as an abstraction, associated with job loss and slop — and behave toward a specific tool that solves their problem. Both are real; they simply are not about the same object. Several other items in this digest are instances of it.
What's the community saying?
The prediction resonated because this digest contains the evidence: an HN thread mourning AI slop, another arguing over whether HN itself is machine-written, a shop advertising that it uses no AI at all — and simultaneously a 13,500-like tweet about a productivity skill and detailed arguments about which coding agent to switch to.