I’m @froztbyte more or less everywhere that matters

  • 8 Posts
  • 654 Comments
Joined 1 year ago
cake
Cake day: July 2nd, 2023

help-circle




  • a long AI sneer on gabi belle’s channel

    the comments also have some bangers. a quick selection:

    My mom used to transcribe calls for companies as work, but then her work implemented AI to transcribe the calls, leading to work calls having sentences in them such as: “Thanks for watching!” “This venue has the best beat” and putting in random websites that DONT EXIST into the calls, when the caller said nothing like that.

    imagine this in financial or medical services

    had a survey from my university about how staff feel about AI implementation for our university. One question that concerned me was “how do you feel about AI being used for grading?”

    The issue with AI grading assignments is that students will most likely get the mindset that if the professor or TA doesn’t care enough to put in the effort to look at my work to give me feedback to do better, why the hell should I put in the effort and spend thousands to do this?

    please come to our university, it costs 500k for a year and we pinky promise that the prof/TA won’t get more than 10% of it, your learning is our “top priority”

    My friend and I entered an art contest a few months ago and she lost to very obvious ai “art”. She spent months on a self portrait oil painting that even when putting my bias aside was amazingly beautiful and definitely deserved to win. There was ai competing in the charcoal category with me too but luckily someone looked at competing pieces before the judges actually scored anything and bombarded the hosts of the competition to remove the piece and they did eventually. It’s incredibly frustrating especially considering part of the first place prize was a scholarship that my friend definitely deserves and needs, and the contest hosts were very hesitant to remove the ai “art”.

    it’s weird how a chunk the art world continues being bad at handling fakes

    For additional context, training ChatGPT-3 took enough energy to propel the titanic at full speed for 37 hours. Or around 920 tons of coal.

    just absolutely mental