So, here’s an example of how AI might affect expectations around search –
Yesterday morning, had a running discussion over Snapchat about good music and new songs with my daughters. Afterwards, I wanted to listen to everything we talked about, so I just copied the entire thread and pasted it into Spotify’s search, assuming (stupidly) that it would just figure it out.
Needless to say, Spotify did not “figure it out.”
(I actually spent 30 seconds looking for that icon of the little floating diamonds, which I suppose has become the accepted shorthand for “We have AI features now!”)
Then I pasted it into ChatGPT and said, “Give me Spotify links for all the songs mentioned in this thread.” It did a nice job. I had to click on them all separately, and they came up in a browser (not the desktop app), but it was better than looking through the text, finding the song names, and searching for them all separately.
Much like Google reset our expectations around search, AI is probably going to do the same thing. Remember when “fuzzy” search was amazing? When Google could correct our typos and search for what we really meant? Pervasive AI interpretation is probably going to make us (or a new generation) assume, “Doesn’t all search just work like that?”
(I remember a video where someone handed a toddler a physical magazine. The toddler, very accustomed to their iPad, kept trying to “click and drag” their finger across the pages.)
I’m certainly skeptical about the rush to add AI to everything, but parsing and otherwise making sense of vague search queries might be a common enough use case to justify it in a lot of software.