I’ll give just one very small example: I have this huge photo library on my phone and sometimes I need to find a particular photo or receipt from a specific business that I clicked 3-4 years ago. Doing that manually means spending at least 10 mins to 30 mins if I don’t remember the exact month or year I clicked it in. And I used to mostly defer finding the photo as much as I could. Or do it when I was totally bored and nothing to do.
Cue AI: I type the business name or photo description in the search bar and it displays the photo right there. 3 seconds.
Another could be about the Camera app using CV to read live text or numbers. I guess CV and LLMs are specifically going to explode this year as they already have a strong base and proven use case.
Roughly 5 years ago it seemed amazing that you could search through Google Photos for "dog" and find all your dog pictures. That seems like such a minor milestone now, when you compare it to the way you can today generate amazing text & visual content matching a specific dog that you describe with a single sentence.
And i would love for researchers to not have to code