'Omakase' is such a creative service design concept that balances unpredictability & delightfulness. So what if we applied that to leadership presentations?

Hi all, wanna hear folks' thoughts on 'how to find a good problem to answer' for new emerging AI concepts (like GenUI/harness/…). My painpoint is that there is so much to explore, but not all questions are worth the time researching. Some might be temporary issues that may not become issues as models get better; some might not even be 'design problems'. E.g. designers may not be able to explore the 'unpredictability side of GenUI' as the UI templates won't fully address tech limitations. Maybe to explore the use case? Scenario? When error tolerance is low?

'Reading the room' is such an interesting concept. AI can't do it (yet). But that can be where the value comes from. E.g. the Uber driver's app playing 'plz use your seatbelt' when the human driver doesn't want to remind the passenger. E.g. a note-taking AI pointing out 'ppl spent most of the time discussing but never actually reached alignment.'