@NathanMerrill My favorite is either metalcore (Architects, Silent Planet, Northlane) or emo/emo-rock (Brand New, Sorority noise, basement, tiny moving parts). But I'll listen to a bunch of other stuff too. Like techno (no particular artists, but stuff on YT like trap nation), blues rock (royal blood, grandson), or a variety of other things
There's another genre that I really like but I'm not exactly sure what the word for it is. Stuff like Walk the moon, spoon, young the giant, and nothing but thieves, I guess?
yeah, midi tracks can make it very easy, but what you need is computer input of when the person is playing. If you notice, there are actual lights on his piano, indicating that it is an electric piano
@DJMcMayhem yeah, but I don't like her most popular stuff. Except for King of Anything.
I shouldn't say that I don't like it: I've just heard it too many times, and it's still my least favorite musically
@NathanMerrill I'd say qute diverse I probably know the old well known composers better than anything newer but it is a little bit hard to say where you draw the line:)
@flawr Where would you recommend starting with these? I think it's the funny that we both listen to a lot of "Metal" but our interests in it are totally different
About a year ago I discovered quite a few french metal bands
But I also discovered youtube channels like worldhaspostrock that just feature an endless collection of post rock, which I now quite like for listening to at work
Mainly because it seems that post rock is rarely with vocals it is not so distracting
But I also really like browsing sites like bandcamp to find some really unknown indie artists
@NathanMerrill I liked these. I didn't really care for the Sara Bareilles song, but the other two were great. I remember you recommending William Joseph a long time ago. And he has a cover of a Muse song, which I really like (Butterflies and Hurricanes).
Have you ever listened to Two Steps From Hell/Thomas Bergersen? He's my wife's favorite composer
I have mentioned this before, but the game in Twin Tin Bots. Premise: you organize tokens to make programs for your two robots. Your robots roam around on a board, trying to collect "crystals" / stealing from other bots. The catch is that you are only allowed to edit one token of one robot per turn, but both robot's programs are executed each turn.
Right now I have most of the game logic implemented... enough to allow scoring and thus develop some AI heuristics for scoring. My game tree code can visit about 100,000 nodes/second.
The issue is that there's a few game mechanics which make things a bit hard. The main one is that, on rare occasions, a player can make two edits in a turn. That bumps the legal move count from 50 to like 1000.
And it's a 2-6 player game, so you can't even thoroughly brute-force one whole round of the game.
So I'm trying to think of some good heuristics.
And also how to combine heuristics / evaluations from multiple heuristics, so that the most meaning can be extracted from the game tree in a certain amount of time.
I think my goal would be for the AI to "make a plan" that's 7 or 8 moves deep.
So 20 ply minimum, but (in my dreams) even up to 50 ply in the worst case.