Would this be along with T154747: FileAnnotations prototype: Search: Guide user during empty state? Technically, I think making smart suggestions that are actually helpful is going to be challenging and require a lot of testing (similar to the way Discovery team is doing it). Do you think this is worth testing in the prototype right now?

In T154754#2927386, @Prtksxna wrote:

Would this be along with T154747: FileAnnotations prototype: Search: Guide user during empty state? Technically, I think making smart suggestions that are actually helpful is going to be challenging and require a lot of testing (similar to the way Discovery team is doing it). Do you think this is worth testing in the prototype right now?

I'd put it in a different way: making smart suggestions requires development effort, that effort is only worth it if these suggestions add value. Simulating those suggestions in the prototype can help us understand how much value they provide. So, if we think the idea is promising, I think it would be worth identify a hypothesis (e.g., it is convenient to link images you uploaded recently), illustrate the idea in the prototype, and include questions in the testing plan to validate the hypothesis.

Hmm. This is too vast and might complicate things. Whether or not a test with smart suggestions succeeds or fails might depend on the smartness of the suggestions, whether they are recent images, from the same category, or from some crazy ML. Lets keep this out if scope of the prototype testing, for now.