When better UX takes out the human

Ethan Mollick points to “the end of detailed summarized thinking traces for Claude and ChatGPT”. As a UX instructor, I think this might be happening to simplify the user experience. Ethan shares that view and he notes it is “likely done both to prevent distillation and to simplify the user experience”.

This “has made being a human in the loop challenging,” says Mollick. “It pushes us towards automation, rather than augmentation”.

It has become much harder to weigh in on in-process Al work.Yes, this is because thinking traces have gotten too long (and too proprietary) but the solution of hiding them almost entirely is a problem. Now, the Al justdoes “agic” and gives you outputs. This feels like a failure of imagination and a move towards a default of disintermediation that is worth challenging rather than blindly accepting.

I teach hundreds of business owners how to enhance productivity with A.I. I am trying to show people how to set up intervention points in their usage of AI because by doing that you can catch wrong assumptions. This means having a plan, pointing to trusted sources, and reviewing outputs before committing to actions.

Reviewable artefacts beat readable monologue. (Thanking Abhjit Ghosh for his insights.)