Skip to content

Comment on I'll never forget how badly the term "AI" nosedived into being uncool

You are viewing one comment and its replies. View all comments · View the comment above

Comments

  • AI is the broader term. LLM is only one kind.

    36
    • Except LLMs aren’t AI. There’s zero intelligence. We should normalize not calling things that lack intelligence AI.

      22
      • You may be using a different definition of “intelligence” than artificial intelligence researchers do

        38
      • There is intelligence though, just not most of the pieces we associate with human like intelligence.

        As a precursor: a simple air conditioner has enough capability to fit most criteria. A basic history and a clock makes is fit a lot more.
        I’m not arguing that LLM are more than they are, rather that the floor is lower than most people think.

        The air conditioner can measure the value of a probe, compare it to a set point, and control a compressor with the result. Give it a basic memory and a clock and it can start making changes before the difference grows. A little more and we can make it improve the quality of that correction over time.
        This is a smart system. It acts intelligently. It has no sapience, reasoning, or any qualities we would associate with a smart creature.

        Intelligence in the context of AI purely means that the system handles tasks previously relegated to intelligent beings, not that the system is itself a being of any sort.
        Things like spell check are early AI. Measuring a set of values, comparing to a table of knowns and picking a weighted blend is another, like you find in rice cookers and washing machines.
        You don’t have to think as hard because it’s taking care of it for you.

        The LLM is predicting the next symbol in a sequence. It’s able to feed that back into itself to produce longer sequences, and can use that sequence to save portions of the sequence to propagate to later inputs to better adapt the next sequence of generated content.
        It’s not reasoning, or feeling, or even thinking. But none of those are required for a system to be intelligent.

        13
        • Exactly. Even OCR used to be AI many decades ago. Nowadays, you can grab text from a photo using OCR, but this isn’t the kind of AI people love to complain about. When the camera finds a face and automatically focuses on it, that’s AI again.

          The floor is very low indeed. A thermostat isn’t intelligent, but many features in a modern one are.

          6
      • Yeah, but none of it is intelligence. It’s an approximation of artificial intelligence.

        11
      • Except LLMs aren’t AI. There’s zero intelligence.

        Almost as if the “intelligence” is “artificial”

        2
        • Artificial intelligence would still be intelligence. LLMs aren’t (and it’s arguable to what extent they’re artificial, given that the models aren’t built or designed, but trained on stolen pre-existing data).

          0
      • I agree with your semantic technicality. AI is properly a human study in the vein of parapsychology and metaphysics in that it doesn’t exist. However in common usage AI is the broadest term encompassing ml llms and other functions.

        2
    • No AI is a made up term for statistic models tech Bros used to “train”.

      There’s machine learning and a subsidiary of ML are LLMs, same as LDA, Random Forest, Decision Trees and other stochastic models.

      The concept of those models isnt even remotely new, the first neural models were hypothesized and trained in the 60’s. Although with only a fraction of the layers of today’s models, since compute power wasn’t remotely close to what it is today.

      These models only can exist because we (humanity) got so good at building tiny transistors.

      And as a studied Data Scientist it boils my blood whenever some problem is supposed to be solved with “let’s use ai for it”.

      18
      • I miss and yearn for the world where I can work hard at something, study it, and have it be worth a damn without it being obfuscated by heaps of desperate and powerful lies. The last 4 years or so have been society melting down, or perhaps “building up” into this weird, fucked up, magic wizard tower, house of cards where nothing is real anymore and nothing is reliable.

        Has the world that is reliable and real ever been there? Or am I imagining that? Was that just a figment of my hopefulness of my youth? I feel like it was real up until basically 2020. Then things started getting really weird and started taking turns and being freakishly hard to predict… Then got easy to predict in the worst way…

        3
        • I mean politically it has turned way worse in some countries. But if you look at it historically, it’s not the first time the super rich have a stranglehold on most of the society. I also don’t believe that they’ll be able to hold on to it, since (let’s be honest) at least the new American super rich are complete idiots. I also don’t think that the “ai” companies of today will survive as they do at the moment, this will and has to come crashing down again. And as my mom said: “regime change begins at home”.

          1
    • What would be some other kinds of modern AI that deserve hate alongside LLM’s?

      3

Sign in to comment. Sign in