Login
You're viewing the functional.cafe public feed.
  • Nov 16, 2026, 11:54 AM
    "If I'm forced to disclose I use AI, people will stop using my stuff :((("

    Take the hint then, read the goddamn room. Either do your slop and stand by it or stop, easy as that. And if you want to lie to people about what you do, too bad, shows really how you are as a person.
    💬 5🔄 61⭐ 0

Replies

  • Nov 16, 2026, 2:46 PM

    @jae Found this exact attitude from a game dev of a game (FortressCraft) I was looking forward to playing :/ Guess I won't then.

    They said they don't want to disclose their usage on steam (despite it being the guidelines to do it).

    💬 1🔄 1⭐ 6
  • Nov 16, 2026, 4:56 PM

    @rubydragon iirc, steam carves an explicit expection for preproduction and code. Code is constantly being treated as some thing seperate from the art of making a game, and it is driving me bonkers.

    💬 2🔄 0⭐ 7
  • Nov 16, 2026, 5:18 PM

    @and i'm not 100% sure on the details but i'm fairly sure they were generating assets and sound/music too, not only code

    💬 1🔄 0⭐ 1
  • Nov 16, 2026, 5:28 PM

    @rubydragon Oh, then never mind then. They be Prim and Proper headassery instead of Following The Rules headassery.

    💬 0🔄 0⭐ 1
  • Nov 16, 2026, 8:20 PM

    @and @rubydragon Me too but there's some twisted logic to it.

    Intelligence is the ability to solve novel problems. When they call LLMs intelligence, they really are trying to say that LLMs are good at solving problems they haven't seen before. Which is absolute BS when the matmul plagiarism machines are trained on the entire internet and then some.

    But creating art doesn't take that kind of intelligence. Look how shit the models are at images with text in them.

    💬 2🔄 1⭐ 2
  • Nov 16, 2026, 8:22 PM

    @and @rubydragon The large multimedia models (LMMs? does anybody call them that?) reveal how the shit works - they plagiarize much more visibly, just like older LLMs generated verbatim stolen code much more often.

    Eventually, they'll probably figure out how to mix up the vectors enough. Just like LLMs got better due to being able to detect and autocomplete more complex patterns and as a bonus (for them), the output was harder to link to input.

    But the underlying mechanism is still stealing ofc

    💬 0🔄 0⭐ 0
  • Nov 16, 2026, 9:36 PM

    @martin_t @and @rubydragon AI has a potentially good role, but it does not substitute for creativity, and when it is used by $$$ hungry people to wholesale disenfranchise others then it becomes a challenge that capitalism can no longer ignore, and likely can’t deal with on it’s own without socialistic oversight, making whatever decisions need to be made. Even the capitalists should have thought about what happens when my markets are disenfranchised? 🤔

    💬 0🔄 0⭐ 0
  • Nov 16, 2026, 4:52 PM

    @jae The bar of expectations from AI people is so low. At this point it is just "please leave me behind and stop lying to me".

    💬 0🔄 0⭐ 4
  • 💬 0🔄 0⭐ 1
  • Nov 16, 2026, 8:55 PM

    @jae “If I’m forced to be honest about what I’m doing then I’ll have to face consequences and that’s just not fair” is a wild position for an adult to take.

    💬 1🔄 0⭐ 8
  • 💬 0🔄 0⭐ 2
  • 💬 0🔄 0⭐ 0