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

help-circle
  • This isn’t a new thing but I hate anytime it asks me a question. I’ll be driving through an accident scene trying to work out where the cop directing traffic wants me to go and if I’ll need to go a different way because the turn I was gonna make is blocked off and at that precise moment google maps decides it’s a great idea to cover the bottom half of the screen with a “is tHeRe sTiLl An aCcIdEnT hErE?”

    If it’s illegal to use your phone while driving it should be illegal for navigation apps to suddenly require interaction in the middle of navigating.












  • I watched their video about this and I don’t think it’s really that simple. Yes, approval voting means your first choice will never harm your second choice, but your 2nd choice can still absolutely harm your 1st choice. In a situation where there are 3 (or more) approximately tied candidates, voting for any candidate which is not your favorite is likely to harm your favorite’s chances.

    Also, their example of a “spoiler” effect doesn’t really convince me. Sure, they negatively frame it by calling the winner party “bad”, but that candidate got more first round votes than the other candidates so it’s logical for it to have an advantage over a candidate who is mostly a second choice.

    It doesn’t seem like a likely scenario either, it requires a scenario where voters who prefer candidate A want B as their second choice, but voters who want candidate B have no agreement with candidate A.


  • Asifall@lemmy.worldtoTechnology@lemmy.world...
    link
    fedilink
    English
    arrow-up
    21
    ·
    1 year ago

    The mentioned but unsupported link to “general intelligence” reeks of bullshit to me. I don’t doubt a modified LLM (maybe an unmodified one as well) can beat lossless compression algorithms, but I doubt that’s very useful or impressive when you account for the model size and speed.

    If you allow the model to be really huge in comparison to the input data it’s hard to prove you haven’t just memorized the training set.