What's this about?
Responding to Willison's ChatGPT Work explainer, Mollick praises the guide and then makes the complaint: the AI labs need to stop making people like Simon — and, to a lesser extent, him — be the ones who explain the modes they release. "Come on, lab folks, you have an AI that can write documentation! And make explainers! And be a tutor! You can do this!"
Why don't the labs write it?
Partly because documentation is unglamorous and ships after the launch. Partly because these products change weekly, so anything written is immediately stale. And partly, less charitably, because the community writes it for free and does a better job — an unpaid explanation from a trusted independent is worth more than an official one, so the incentive to fix it never quite arrives.
What's the community saying?
966 likes — one of his most-liked posts of the window
- The irony is the whole joke and it lands: the product being undocumented is a text generator.
- It is also a labour point. Two people are doing the field's documentation unpaid, and the engagement they receive is not compensation.
- Willison's tool reference, published the next day, is the systematic version of the same unpaid work.