• 0 Posts
  • 46 Comments
Joined 1 year ago
cake
Cake day: June 12th, 2023

help-circle






  • By your definition of harm, no artist creating non-material goods (books, movies, music, etc) could ever experience harm due to any one individual’s actions. “I was never going to pay, so taking it without paying is a victim less crime,” etc, etc.

    False. I acknowledge that there could be harm if a consumer would otherwise be able to afford to pay for all of the music they listen to. The distinction here is that if a consumer is already spending as much as they can truly afford then artists aren’t going to get any more money out of this consumer, regardless of whether or not they pay for it.

    In other words: if you pirate because you must = no harm; if you pirate because you can = some harm.

    That’s an interesting thought experiment about the cheating spouse, though. Thank you for the interesting perspective! This makes me want to re-visit my philosophy notes.

    For the record, I pay for Spotify and also support artists through Bandcamp, merch, vinyl, and live concerts. I also pirate music which isn’t otherwise available through Spotify and/or Bandcamp (e.g. The Grey Album by Danger Mouse, and up until recently The Flamingo Trigger by Foxy Shazam) and don’t feel guilty about those instances.









  • Good question. Neural networks are modelled after how brains learn and process information, so it’s certainly theoretically possible for a neural network (or other machine learning algorithm) to make inferences like that, just like how you’ve learned them with years of experience.

    The biggest challenge in any machine learning is finding enough labelled training data. In fact, a friend of mine contributed to a paper in which (no joke) GTA V was used to generate labelled training data for an automous vehicle. Because it’s a game engine, every object in the game is already digitized, and the 3D modelling is accurate enough to be useful, at least. This vehicle used LIDAR so the actual shaders and such didn’t matter as much as the 3D point cloud.