What's this about?
Willison wrote up what ChatGPT Work can actually do, describing it as "a deeply confusing but extremely powerful tool with a whole lot of useful features that aren't available in regular ChatGPT." This follows his question a few days earlier asking users when they switch to Work — a question nobody could answer cleanly.
Why does a product need a third-party explainer?
Because the feature set and the affordance have come apart. Work has capabilities regular ChatGPT does not, but the interface does not communicate which task belongs in which mode, so users default to the familiar one and never discover the rest. When someone who tracks this space professionally has to write the missing documentation, the gap is in the product, not the user.
What's the community saying?
531 likes on the writeup, 222 on the original question
- The original question — when do you use Chat and when do you use Work — produced no consensus answer from experienced users, which is the finding.
- Willison's framing accepts both halves: genuinely powerful, genuinely confusing, and the confusion is not the user's fault.
- It rhymes with the model-routing debate: whenever a system asks the user to pick which capability handles a request, most users pick wrong or pick once and never revisit.