156
Nexus Mods now requires creators to disclose AI use but isn't ready to ban AI-generated mods (www.tweaktown.com)
submitted 3 days ago by to c/pcgaming@lemmy.ca
22 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 1 point 2 days ago* (last edited 2 days ago) (10 children)

Indie dev here and day job is an AI consultant, engineer and academic. Truth be told genai just isnt that good. It averages to the mean so it will only really be useful for some things. The issue isnt no AI or all AI but about disclosing how much and when. Using llms to support (but not replace) programming borrows from stack overflow so is less a copyright issue. But fully replacing your code or artists with gen ai will make for shit games.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 10 child comments
    •
  • [–] -3 points 2 days ago (9 children)

    I haven't talked to anybody who's replacing IT, engineers, or developers, but let's be real programmers are done. Even sysadmins will outlive them.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 9 child comments
    •
  • [–] 2 points 2 days ago (3 children)

    Said by someone who, presumably, has never tried to program something technical because LLMs are terrible at it.

    I'm using AI quite a bit to help me write code, for work, but I need to review its output so carefully. While cheerfully reporting that it was operating on a test instance, and that no destructive operations would touch live servers, it never toggled to use the test server variables at all. Even worse, it had hard coded using the live server, but only for the delete operations. That would have been a disaster if I was slop coding it, instead of using it as an AI assistant, and very carefully reviewing every line it wrote.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 1 point 23 hours ago (1 child)

    Hard disagree. LLMs I simply next word predictors that's all they really do. It's the human driving them that makes the decisions.

    LLMs don't "build" they simply do what they're told. So they're only as good as the engineer using them.

    I think you're highly overrating them.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 22 hours ago

    Not sure how you got that from my comment... I said that it's completely unreliable and can't be trusted to do even basic tasks, like not deleting the live server (in my example).

    I'm guessing I didn't explain that well in my advice post? But it sounds like I mostly agree with you.

    (I think saying they "do what they're told" is a bit too strong of a statement, but I think I get what you're saying.)

  • permalink
  • fedilink
  • source
  • parent
  • [–] 0 points 1 day ago

    Most fail to grasp LLMs were trained on over confident stackoverflow posts and terribly lazy tutorials, and worst of all we turned off human to human feedback loops to give it more training.

    Its why comp sci circles of mine joke/worry we cant get past python 3.12 because LLMs have no human-created training data for new versions now that we stopped using traditional human to human platforms for that.

  • permalink
  • fedilink
  • source
  • parent
  • [–] -1 points 1 day ago* (last edited 1 day ago) (2 children)

    Lol. No. I am a life long programmer and founded a successful, profitable AI company, and train governments and industry around the world (my name is here, you can find me to verify) and I think this is hogwash.

    Programming depends on creative solutions to new problems while LLMs are trained on an increasingly diminishing source material. They cant move past python 3.12 is the joke/criticism in computer science circles.

    Evidence shows llms make engineers SLOWER and any serious engineer who gave them a full try understands why. They guess poorly and confidently, erase logic, have poor architecture.

    https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/

    https://fortune.com/2026/03/10/ai-productivity-workers-workday-efficiency/

    LLMs are pattern machines not sources of truth or intelligence.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 0 points 23 hours ago (1 child)

    I think you're really overestimating these LLMs.

    All it is, is a next word predictor. It's not intelligent and it doesn't make decisions. It's only as good as the engine you're driving it.

    Nah, you overestimate these tools.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •