936
Surely they yearn for the mines (thelemmy.club)
submitted 1 day ago by to c/lemmyshitpost@lemmy.world
91 comments fedilink hide all child comments
 
you are viewing a single comment's thread
view the rest of the comments
[–] 9 points 16 hours ago (2 children)

The very premise of this meme is ridiculous.

Mining was one of the earliest adopters of automation. "Robots" have been doing mining jobs for decades, for centuries if you are generous with how you define "robot". Mining companies keep searching for any job they can automate. Even the guys who drive the enormous mine trucks were replaced when they found they could let a machine drive the trucks instead.

The only difference is that nobody yet knew how to automate "knowledge work". The latest generation of AI can simulate knowledge work, which apparently is good enough for some jobs. But, the fact that it hasn't yet taken some of the few remaining mining and chemistry jobs is a clue that it has severe limitations.

Sure, you can automate the creation of powerpoint presentations, which is "knowledge work", but is there a mining company that is willing to fire the guy who supervises the robots and replace him with ChatGPT? These are ruthless profit-driven companies. If AI could do that guy's job, he'd be unemployed.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 3 points 15 hours ago (1 child)

    Current LLMs are limited by the fact that they can't train during inference. i.e. They can't learn. If we find something better than backpropagation or diffusion, then things are going to actually be different. At that point, the models would be able to compose ideas together instead of just predict the next token in sequence based on previous training.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 14 hours ago

    Current LLMs are limited by the fact that they can’t train during inference. i.e. They can’t learn

    Learning is much more than just training during inference.

    True unsupervised learning is still in its infancy, and requires a completely different architecture than the current fancy auto-suggest boom. Unsupervised learning has been the dream since at least the 1950s when AI was an aspiration, even before transistors existed. Nobody has figured out how to do it yet.

    If pigs learn to fly, then things are going to actually be different.

  • permalink
  • fedilink
  • source
  • parent