A classic nerd from Norway.

  • 0 Posts
  • 28 Comments
Joined 1 year ago
cake
Cake day: June 15th, 2023

help-circle


  • Impersonal social media like reddit, lemmy, twitter, and mastodon is nice sometimes. We can have a conversation without the conventions, responsibilities, and limitations that comes with real life socializing.

    I can decide for myself how long I want to think of a reply. And a reply isn’t even expected, I can choose not to respond if something makes me uncomfortable to talk about. Nor do I have to wait my turn to converse, or be spoken over. I dont have to remember faces or names. Nor remember peoples background or recent events or drama they’ve been involved in. Theres no judgment on age, looks, ability, or gender (unless I disclose it of course). And its possible to fact check stuff while discussing, no need to wait until after to realize someone is trying to misinform.

    Imho, all this turns toxic when companies involve “recommended feeds” like they often do.


  • This is probably very unlikely and I got no idea what I’m talking about: But what if feeding it even small amounts of its own content, text produced by a chatgpt instance, poisons it? That it gets confused from being fed text that adheres perfectly to its own rules, and locks that text down as perfect and not needing small variations.

    I remember some article warning about this in a big scale, and I’m thinking why must it be big? If its only a probability tree, even small changes to the probability would cause issues further up the branches.

    But blind speculation.