• 0 Posts
  • 22 Comments
Joined 1 year ago
cake
Cake day: June 11th, 2023

help-circle
  • They’re supposed to be good a transformation tasks. Language translation, create x in the style of y, replicate a pattern, etc. LLMs are outstandingly good at language transformer tasks.

    Using an llm as a fact generating chatbot is actually a misuse. But they were trained on such a large dataset and have such a large number of parameters (175 billion!?) that they passably perform in that role… which is, at its core, to fill in a call+response pattern in a conversation.

    At a fundamental level it will never ever generate factually correct answers 100% of the time. That it generates correct answers > 50% of the time is actually quite a marvel.





  • ApexHunter@lemmy.mltoProgrammer Humor@lemmy.mlcoding chess
    link
    fedilink
    arrow-up
    7
    ·
    edit-2
    9 months ago

    The reply would have been return x % 2 == 0, or if you wanted it to be less readable return !(x&1).

    But if you were going for a way that is subtly awful or expensive, just do a regex match on “[02468]$”. You don’t get a stack overflow with larger numbers but I struggle to think of a plausible bit of code that consumes more unnnecessary cycles than that…




  • Agree no difference as an ingredient in some baked dish.

    But if you are eating the egg by itself or as the primary item, there is definitely a difference in taste. Not a revolutionary change your life difference, but still a difference.

    In my experience the difference is pretty small amongst the options in the grocery store, but fairly noticable for eggs I get from the farmers market.






  • For first names we separately compiled lists of names that we liked from whatever sources we could find (Internet lists, books, media l names used in media, etc). We went through each other’s lists and vetoed names that were a hard no. Then we wrote the names out on a sheet of paper in random order in a playoff bracket style arrangement. Each pair had a winner until there was only one.

    For middle names, it had to be something that flowed well with the first name. It also had to be able to convey that special sense of “you done fucked up” and disappointment when said with the first name (while emphasizing the second), like “John PAUL”. Finally, candidates were from (mostly deceased) family members.

    Names and initials were checked to minimize bullying potential; if we could think of a way to abuse it the name or combination was rejected. For example, Karen would be a no due to current slang usage. Or if the initials would spell ASS.