Jump to content

Slowswift

+Patrons
  • Posts

    3684
  • Joined

  • Last visited

  • Days Won

    1

Slowswift last won the day on July 23 2024

Slowswift had the most liked content!

About Slowswift

  • Birthday March 14

Profile Information

  • Member Title
    π―π‘Š 𐑅𐐴 𐐿𐐫𐑍𐑉𐐭
  • Location
    Our Lovely Deseret
  • Interests
    Reading (duh), writing (theoretically), pens, flags, music (alternative, electronic, and classical mostly. K- and J-pop also welcome), languages and scripts, trains, astronomy, anime, assorted fandoms.

Slowswift's Achievements

4.8k

Reputation

Single Status Update

See all updates by Slowswift

  1. Ignore the provocative title (or don't, I'm not your mother), but I think this is the video that's finally radicalized me against AI. Not to say that I just adopted her opinion wholesale--but rather, she distilled all the various thoughts I've been having into one place.

    I've used AI very sparingly. From the outset, I've refused the idea of generating prose with it, or even using it to come up with an outline or generate ideas. Rather, I've used it like a kind of rubber duck that can talk back. I explained my premise, then asked it for questions to prompt development. This worked great, for a little while. But the further the conversation went, once it started editorializing, the problems started appearing. There were hallucinations and straight-up incorrect information, of course, and several times it took a passing idea of mine as gospel. I'm abundantly aware that these things are programmed to largely just tell you what you want to hear. The time I asked it for feedback on a scene was illuminating: maybe one of its suggestions was good? The others were technically valid observations, but I elected to not take its advice on the grounds of creative differences. And don't get me started on the tone of voice; that vapid LinkedIn cadence. (You can alter that a little bit, and I eventually got it to a point where I could tolerate its commentary, but the point still stands.)

    I've played around with image generation too, because on a technical level I do find it interesting, but aside from the ethical concerns it's just not very good--and the effort needed to prompt it into making something good vastly outweighed any inclination to try. Old-fashioned mood boards and faceclaims it is, then.

    In the end, as useful as I've found it to be, I'll be greatly winding down my use of commercial chatbots. (I'll still keep an eye on the technology, because I do find it fascinating, and at this point the genie's out of the bottle.) Worry for my mental capacity aside, I just don't trust that any of the companies behind them have anything remotely approaching my best interests in mind.

    I think I'm going to go read a book, now.

    Β 

×
×
  • Create New...