Last week a Nature Communications paper described scientists using AI to propose new designs for one part of the natural factory bacteria use to make antibiotics and other medicines. They built and tested hundreds of versions in real cells, and the best made up to about three times more of the product than the natural version, so the AI supplied more ideas worth trying and the lab work stayed with people.

The same week Google started rolling out Guided vision in Gemini Live on Android, a camera mode for blind users that describes what’s around you, reads text, and tells you where to point the phone. One early reviewer expects to spend less time reminding Gemini that they’re blind, and deciding where to walk is still the person’s job.

In a Boston University study, online MBA students given an AI tutor that asks questions and waits for their attempt before explaining did better on the end-of-module test than classmates who had ChatGPT and the like freely available, mostly in how well they reasoned in writing, since the tutor left the thinking to them.

All three are the kind of AI I’m most comfortable betting on, where the tool widens what someone can attempt and the judgment stays where it was.