>Me actually trying to solve the neuralink challenge only to realize that it's obviously fricking impossible

>Me actually trying to solve the neuralink challenge only to realize that it's obviously fricking impossible
Was it a prank?

Shopping Cart Returner Shirt $21.68

Unattended Children Pitbull Club Shirt $21.68

Shopping Cart Returner Shirt $21.68

  1. 2 weeks ago
    Anonymous

    >the neuralink challenge
    first time hearing about this, what is it

    • 2 weeks ago
      Anonymous

      https://content.neuralink.com/compression-challenge/README.html

      • 2 weeks ago
        Anonymous

        It would be feasible if it allowed some loss of quality (or not idk im talking out of my ass)
        Why would he want lossless compression anyways? They are recording movements from muscles, they do not require to be exactly precise

        • 2 weeks ago
          Anonymous

          They're recording ambient CSF voltage, I don't know why they need 10 bits of lossless precision.

      • 2 weeks ago
        Anonymous

        >Elon writes a programming challenge asking for impossible compression ratio
        >Is very vague about the requirements: Algo has to "work on Linux"
        >Gives no details about CPU, architecture or even how much RAM can be used
        >Some IQfy mongoloid emails Elon a schizo babble "solution"
        >It doesn't work of course but Elon doesn't care
        >Tweets the "solution", claiming it as his own
        >Claims Neuralink has created a telepathic implant
        >Neuralink shares are up 69420% an hour later

        • 2 weeks ago
          Anonymous

          Neuralink is private, so it's not that, they've probably hit a wall internally and are hoping for some miracle solution with this moronic challenge, in denial that what they're asking for is impossible.

          not to anyone with any experience on the topic. For them it was bait safely ignored. Stop shitting up the board with this Elonshit.

          It's the first and last post I'm making on the matter, but I wasted my Sunday on this and I will vent about it.

          • 2 weeks ago
            Anonymous

            >but I wasted my Sunday on this and I will vent about it.
            It takes like a minute to read that they need 200x lossless compression, laugh at it, read it again, see that they want it realtime and laugh even harder and move on with your day.

        • 2 weeks ago
          Anonymous

          You're right.

          If you solve the challenge, you will get 0 credit.

  2. 2 weeks ago
    Anonymous

    Elon chuds outsmarted yuropoors yet again

    HE CAN'T KEEP GETTING AWAY WITH IT

  3. 2 weeks ago
    Anonymous

    >> 200x compression is needed
    how the actual frick

  4. 2 weeks ago
    Anonymous

    not to anyone with any experience on the topic. For them it was bait safely ignored. Stop shitting up the board with this Elonshit.

  5. 2 weeks ago
    Anonymous

    >fourier transform
    >develop an algorithm to pick what high frequency parts are worth keeping (doesn't have to be perfect at all)
    >chop off the rest of the high frequency data
    wow that was hard
    >its lossy tho
    i don't care

    • 2 weeks ago
      Anonymous

      That or a kalman filter.
      I tried both, both work, not losslessly and definitely not 200 to 1.

      You're right.

      If you solve the challenge, you will get 0 credit.

      If you solve the challenge you should keep the code as leverage and only send a decompression binary with the compressed data.
      If you send the code in no questions asked like a fricking imbecile, you deserve to be robbed.

      >but I wasted my Sunday on this and I will vent about it.
      It takes like a minute to read that they need 200x lossless compression, laugh at it, read it again, see that they want it realtime and laugh even harder and move on with your day.

      Task specific compression can achieve good ratios but this is just a bunch of relatively entropic and probably very noisy voltage measurements, there's no reason to go for lossless here.

      • 2 weeks ago
        Anonymous

        lossless is a moronic requirement
        the sensors are 20000Hz
        they are measuring human neurons that MIGHT fire at like 200Hz

        • 2 weeks ago
          Anonymous

          This was never mentioned, but every electrode seems to have multiple probes, I'm guessing they're cycling through all of them and that's why it's 20khz, though I could be wrong about that.

          >thread brainstorming about what could be done to make a pretty good lossy algorithm
          >Elon can read this for FREE

          I mean if the engineers over there couldn't figure out how to Google "noise reduction algorithms"...

  6. 2 weeks ago
    Anonymous

    don't know a single thing about compression algos
    got down to average ~21:1 compression and gave up once it was pointed out to me that the $1,000,000 I read was promised was, in fact, never promised

  7. 2 weeks ago
    Anonymous

    So from what I am reading so far, the requirements are impossible. That being the case, wouldn't it make more sense to go back and revamp your methodology to they can get a solution that does work rather then hoping for a miracle?

    • 2 weeks ago
      Anonymous

      >Expecting a ELOOON company to do anything correctly

    • 2 weeks ago
      Anonymous

      The only explanation is that whoever wrote that doesn't understand that the requirements are impossible.
      I wouldn't be surprised if it's the man himself.

      • 2 weeks ago
        Anonymous

        >Elon makes some unreasonable demand at neuralink
        >gets told it's impossible
        >fires the people who said it was impossible and tries to get morons on twitter to solve his problem

  8. 2 weeks ago
    Anonymous

    >thread brainstorming about what could be done to make a pretty good lossy algorithm
    >Elon can read this for FREE

  9. 2 weeks ago
    Anonymous

    The FDA already approved a second human trial for the neuralink, so clearly progress in Man-Machine interfaces is chugging along. I wonder how far it will be into the future where such technology overtakes VR.

  10. 2 weeks ago
    Anonymous

    >motor cortex poops out the electrodes anyway
    this homie is so fricking moronic you have no idea

  11. 2 weeks ago
    Anonymous

    >my first time doing multi-threading without system calls, naively thinking that complex data structures were not needed

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