Something I particularly love about AI is its ability to draw patterns out of things that are seemingly disconnected. You can just give a neural net all the information you have in weather patterns and bird migratory patterns and it’ll be like “hey. I can predict when it will rain based on the ratio of the frequency of a bird call and its north-north-east velocity”. I don’t know what sick joke the universe is trying to pull making all these connections (not that this example is real) and I don’t know how someone could ever figure things like that out without AI, but that’s cool as hell and I love it.
Now sometimes this is cool. I’m sure if you’re trying to trade on the stock market being able to tie a bunch of stocks, online sentiment, trading patterns, and the micro second delays of Big Ben’s rings together to squeeze every penny out of your GameStop trades (ayyyy let’s go timely references!) If I ever get some crazy disease and the doctor claims they can cure it using data on my heartbeat, blood pressure, and the exact milliliter volume of my piss, then absolutely let’s do it.
But I swear to god, if a certain music playing AI on a certain music streaming app tries to suggest I listen to my “Thursday-Core” music, I’m going to go ballistic. Maybe Friday or Saturday nights you could make the argument people will listen to different music for parties or to relax or to get high, but not a single soul on this planet has thought “man time for my Wednesdaily Gang of Youths listening to really put a pep in my step!”
There’s so much data you could use to help me find more music. Match up the weird micro-genres of my favorite songs and find more. Match up the frequencies of the singers’ voices or the amount of those lil squeaks from the player moving their hand on the guitar. Those would be interesting points to map. Not once, ever, have I made the conscious decision to have a specific Tuesday soundtrack.
There’s a lot of cool bits of data you can collect and mash together, and there’s a lot of really interesting patterns that can emerge from big data sets, but don’t go assigning meaning to noise to make your music playing robot seem more in depth.
Also all the X is is snippets from the playlists that have existed for years with an occasional TTS thrown in I see through your lies Spotify.