1312
racist ai (lemmy.blahaj.zone)
submitted 8 months ago by to c/fuck_ai@lemmy.world
223 comments fedilink hide all child comments
 

https://marcella.masto.host/@bitchboss/116167880762637310

you are viewing a single comment's thread
view the rest of the comments
[–] 3 points 8 months ago (5 children)

If we weren’t already boiling the oceans, I’d say that you should continue to use the free tier. A lot. Drive up their costs as much as possible without giving them any revenue by constantly asking it to generate useless crap.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] 15 points 8 months ago*

    No, that's a bad idea. They literally make up their economics. They don't care abo about money, because they'll just keep borrowing more and more of it. Just look at their behaviour over the last three years.

    What they really want is usage go up and in turn their power they have over their userbase. You using their product to generate useless crap at best just waists power and water. At worst the AI will give a a psychological complex.

    It won't hurt their bottom line, trust me. You can only loose.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 9 points 8 months ago (3 children)

    It also drives up market share, which is worth valuation to investors.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 2 points 8 months ago (1 child)

    And data collection. People think they’re good if they just don’t give it personal info but you’re definitely being fingerprinted and tracked. If they haven’t already they’ll eventually partner with meta or google (in the case of Gemini that’s obviously already done) to match any fingerprinting to much denser profiling. Plus they’re designed in a way to encourage you to let your guard down and talk casually, which disarms you and makes it more likely you mention identifying info they can’t determine from IP and fingerprinting.

    OpenAI has already added advertisement so this is definitely occurring. No one buys ads in this day and age without robust audience targeting

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 8 months ago

    Well, continual usage doesn't really help with fingerprinting. They already got that.

    I'd say, if you're going to do this (you shouldn't) the prompts should be fed by another LLM, and just let them create garbage together. I'm sure OpenAI likely uses user prompts for training, so this fucks up their training data.

  • permalink
  • fedilink
  • source
  • parent