So let me get this correct

Elon doesn't;
>use Grok
>use OpenAI's GPT
>use Google's Gemini
>use Meta's Llama
>use Anthropc's Claude
to discover an algorithm that losslessly compresses extremely noisy files to 1/200th the size?

this is peak proof that these big tech companies don't know how to make AI as beneficial as they claim or use the tech in any meaningful manner.
What in the grift is going on?
You telling me current AI tech can't easily shit out an algorithm to do this?
if so, then they are scamming the world.

It's All Fucked Shirt $22.14

Black Rifle Cuck Company, Conservative Humor Shirt $21.68

It's All Fucked Shirt $22.14

  1. 4 weeks ago
    Anonymous

    I don't know where you got the impression that any current AI could simply invent an algorithm to losslessly compress to 1/200th original size but it wasn't from any AI company nor anyone that uses their models.

    • 4 weeks ago
      Anonymous

      morons like OP think LLMs are actually intelligent.

      • 4 weeks ago
        Anonymous

        morons like OP don't think. They're NPCs

        • 4 weeks ago
          Anonymous

          this

          you know how stupid you sound? especially when You keeping (You)ing?
          >elon can put brain chips inside of human skulls
          but
          >AI can't find compression algorithms that compresses files to 1/200th the original size
          I really wish they began an IQ test to post on this board because you'd definitely be restricted to /b/, if that.

          • 4 weeks ago
            Anonymous

            [...]
            the beautiful thing is you two jack nuts believe your answers to be accurate.
            you both are like cute idiots, saying dumb shit but we all won't flog you for you because no one likes to see morons being beat, publicly.

            >I really wish they began an IQ test to post on this board because you'd definitely be restricted to /b/, if that.
            Why do you want to restrict yourself from this board?

            You don't even know how your god works. If there's no lossless compression algorithms inside the dataset ad hoc, that the deep learning model has been trained upon, not in a million years would it shit out the correct solution to the problem, it will just hallucinate the answers.

      • 4 weeks ago
        Anonymous

        this

    • 4 weeks ago
      Anonymous

      morons like OP think LLMs are actually intelligent.

      the beautiful thing is you two jack nuts believe your answers to be accurate.
      you both are like cute idiots, saying dumb shit but we all won't flog you for you because no one likes to see morons being beat, publicly.

      • 4 weeks ago
        Anonymous

        >the longest posts
        >the least said

        • 4 weeks ago
          Anonymous

          thanks for bump /b/itch

          • 4 weeks ago
            Anonymous

            Here's another so everyone can see how 'smart' you are.

      • 4 weeks ago
        Anonymous

        Schizo + brown + moronic

  2. 4 weeks ago
    Anonymous

    >doing free work for another one of Musks vaporware projects

    • 4 weeks ago
      Anonymous

      Even if that is the case, why the frick aren't they throwing AI at this problem? I find this to be odd as frick.

      Here's another so everyone can see how 'smart' you are.

      >you right anon
      I know I am, b***h.

      • 4 weeks ago
        Anonymous

        I'm not even the homosexual you're replying to, but you invoked us here by name to tell you that youre a moron.
        hi from /b/, the current ai models will confidentally make up entire apis that dont exist and dont know how to code efficiently. given that every ai is a series of complicated weight maps, how the frick would it ever come up with anything new, moron?
        the less similar something is, the harder it is to compress. please take a cryptography and encryption referesher and keep our board out of your shit covered fingers' type pattern. don't forget where you are

        • 4 weeks ago
          Anonymous

          > I'm not even the homosexual you're replying to
          Then shut your b***h ass up then, speak when spoken to, b***h

  3. 4 weeks ago
    Anonymous

    >You telling me current AI tech can't easily shit out an algorithm to do this?
    Obviously it can't.
    To do that it would actually be intelligent and not just regurgitate a word salad of what thousands of people already did before.
    Calling it AI is even a misnomer. It's not intelligent. A dog is magnitudes upon magnitudes smarter than current LLMs.

  4. 4 weeks ago
    Anonymous

    >What in the grift is going on?
    What in the big chungus is going on?

  5. 4 weeks ago
    Anonymous

    Such an algo exists but it would require a lot of processing power, there's literally always a tradeoff. I wouldn't help Musk though, he's a israeli cuck.

    • 4 weeks ago
      Anonymous

      Shut up

  6. 4 weeks ago
    Anonymous

    This is just some engineers' last ditch attempt at convincing elon that his "just compress it by 200 times bro" idea won't work.

    • 4 weeks ago
      Anonymous

      so one would even say, it's a humiliation ritual.

  7. 4 weeks ago
    Anonymous

    Why lossless? Just fricking filter it a bit, some of that noise has to be completely useless for BCI, the files are sampled at 20KHz, not a medgay but I assume neurons don't spike at ~10KHz? They should also threshold the shit, so many cheap ass things to bring the entropy down and allow for more compression. From wiki it seems like a an action potential takes around 2 ms, meaning you would have to sample at a few KHz to pick it up with high certainty. Why are they oversampling so much and not processing it at all(assuming t=2 ms)

  8. 4 weeks ago
    Anonymous

    It's just a hiring scheme, anyone who calls them out over email and does a pretty good job with compression will probably be hired if they want the job. You can easily sort the poos out like this, they will actually think its possible.

  9. 4 weeks ago
    Anonymous

    >elon musk discovers the show silicon valley
    I think it is probably a meme to prescreen hiring. dial the sampling rate back to 200hz which is enough for their application then hit it with a standard compression will give you a 200x reduction. kind of like a mitwit filter.

  10. 4 weeks ago
    Anonymous

    OP is a moron, but he accidentally has raised an interesting point though.

    We know LLMs are not actually intelligent.
    But normies actually believe they are.
    Thus the massive tech bubble in the stock market which comes from this misconception OP is regurgitating in his posts. Will be interesting to see what happens when stock market expectations meet reality.

Your email address will not be published. Required fields are marked *