I had the idea to try and replace it as my “personal assistant” for a day after the app for Android released the voice chat for the free tier users. I had an old openAI account so I logged in and tried it out, Bluetooth buds in. Walked the dog, made breakfast, basically as if I were on the phone or “using Jarvis”. It was pretty cool, it misheard and repeated a lot of things when I didn’t say anything but for the purposes of when I did use it, it actually worked and was cool.

I’m not a meal or recipe planner but on my walk with my dog it gave me some ideas from what I had in my kitchen. I complained about my wet boots and it suggested giving them a coat of something to waterproof fabrics (like scotchguard). Had a conversation with an AI lol, it was interesting. There were some other things but they’re escaping me.

When I got back I thought to try asking it to mealplan for the week. Initially it just gave me a blanket answer with no details from me and I tried “again but with no ___” and it did pretty well, and it doesn’t read it out for 10 mins it just sends a list in text which is good.

A while back I had the idea to streamline information about songs, list key, root notes, the tempo, the scale used in the song, and scales that can be used along with it. Took a bit of prompt engineering but I got it replicable by just going “now tell me about ____ by ____”. And it had some pretty obscure stuff (like Xetrovoid, a fairly small artist). I’m fairly fluent and they seemed right but I double checked online and they were all accurate - save for the scales for a couple of them.

I don’t think I’d use it for getting information about all of the scales notes in a key but it may be worth trying.