Comment on We need to stop pretending AI is intelligent

• Blue_Morpho@lemmy.world ⁨1⁩ ⁨day⁩ ago

  Given that the weights in a model are transformed into a set of conditional if statements (GPU or CPU JMP machine code), he’s not technically wrong. (Of course, it’s more than just JMP and JMP represents the entire class of jump commands like JE and JZ. Something needs to act on the results of the TMULs.)

  source

  • tmpod@lemmy.pt ⁨1⁩ ⁨day⁩ ago

    That is not really true. Yes, there are jump instructions being executed when you run interference on a model, but they are in no way related to the model itself. There’s no translation of weights to jumps in transformers and the underlying attention mechanisms.

    I suggest reading …wikipedia.org/…/Transformer_(deep_learning_archi…

    source

    • Blue_Morpho@lemmy.world ⁨1⁩ ⁨day⁩ ago

      That is not really true. Yes, there are jump instructions being executed when you run interference on a model, but they are in no way related to the model itself.

      The model is data. It needs to be operated on to get information out. That means lots of JMPs.

      If someone said viewing a gif is just a bunch of if-else’s, that’s also true. That the data in the gif isn’t itself a bunch of if-else’s isn’t relevant.

      Executing LLM’S is particularly JMP heavy. It’s why you need massive fast ram because caching doesn’t help them.

      source

      • tmpod@lemmy.pt ⁨1⁩ ⁨day⁩ ago

        You’re correct, but that’s like saying along the lines of manufacturing a car is just bolting and soldering a bunch of stuff. It’s technically true to some degree, but it’s very disingenuous to make such a statement without being ironic. If you’re making these claims, you’re either incompetent or acting in bad faith.

        I think there is a lot wrong with LLMs and how the public at large uses them, and even more so with how companies are developing and promoting them. But to spread misinformation and polute an already overcrowded space with junk is irresponsible at best.

        source
• elbarto777@lemmy.world ⁨1⁩ ⁨day⁩ ago

  I’ll be pedantic, but yeah. It’s all transistors all the way down, and transistors are pretty much chained if/then switches.

  source
• Hotzilla@sopuli.xyz ⁨1⁩ ⁨day⁩ ago

  Calling these new LLM’s just if statements is quite a over simplification. These are technically something that has not existed before, they do enable use cases that previously were impossible to implement.

  This is far from General Intelligence, but there are solutions now to few coding issues that were near impossible 5 years ago

  5 years ago I would have laughed in your face if you came to suggest that can you code a code that summarizes this description that was inputed by user.

  source

  • JcbAzPx@lemmy.world ⁨1⁩ ⁨day⁩ ago

    I think the point is that this is not the path to general intelligence. This is more like cheating on the Turing test.

    source