To this point, I want to rant a bit about the experience that most convinced me that they actively hated functionality. So I use Pandora for music, just seemed to feel the best. It’s not the biggest player out there, not as much direct integration as something like Spotify, but I have simple needs. I could just tell Assistant to ‘play Pandora’ and the app would open on the one station I use and get going.

Well, Gemini rolls around, and it even says it can fall back to Assistant if needed, so no reason not to try it, right? Of course, if you ask Gemini to ‘play Pandora’, all it does is tell you “I can’t use Pandora yet. Try YouTube Music or Spotify.” or similar. How hard can it be to make it understand the user wants to open this mainstream music streaming app and hit the media play button? Too hard for an AI engineer, I guess. Oh well, if you tell Gemini to ‘open Pandora’, at least it will open up the app–after you manually unlock the phone.

Side note, the voice match is already so selective it only gets me like a third of the time (unless I’m reading a crossword clue from across the room), and I’m not exactly jamming out to a list of all my passwords read out over some looping beat, I think you can just go for it. I swear it used it to be better at detecting voice and snappier about responding when it did, but I guess these giant software companies have sped up devolpment so much it’s actually rolled around and started going backwards.

Anyway, ‘open Pandora’ at least does something, though I’ve already gotten into a habit of saying ‘play’ instead of ‘open’, so it can be annoying at times. I know!, there’s that feature to alias a trigger phrase to a routine of commands. I figure out how to get that set up to turn ‘play Pandora’ into ‘open Pandora’ (which IIRC was harder than it should have been for some reason). Alright, I go to test it out and “I can’t use Pandora …”!!! I literally went in to manually set up a workaround for this issue that shouldn’t even have been an issue in the first place, but do you know what’s more important than actually doing what the user wants? Telling the user how absolute garbage our product is!!!

Alright, time to dump this in the trash where it belongs and go back to the old Assistant like Google promised you could. Sike! You forgot that was a Google Promise™! The old Assistant just hangs immediately and crashes. That’s actually really convenient though, because my headphones have a button you can use to directly activate a digital assistant on your phone, but it can only be set in the mfr settings app to use Alexa or Google Assistant. Since I’m not gonna use Alexa, I can just leave it set on Google Assistant and–wait, did I slip and say “really convenient” earlier? I guess what I meant to say was “a huge pain since Sony isn’t going to correct their short-sighted buffoonery on their (aging, but still going strong) flagship headphones any time soon”. Funny how you can just miss a couple keys and completely change what you meant to say.

I have no idea what Google is thinking with their product development team, but they’ve clearly learned that the best thing to do with an established, functional tool with a good userbase is to just toss it all out the window for the next piece of trash they can put a clever name on. So the question is: Would it be unethical to round up everyone who supports pulling the sort of nonsense I described and “re-educate” them until they reach some baseline level of rationality, or should it go past just non-counterproductive until they can make actually good decisions?