yudkowsky tried to warn you.

yudkowsky tried to warn you. you can keep blowing off AI all you want but it will continue to take over certain industries. and some day soon it will take over all industries because that's what trends show - that's technological progress.

plan accordingly. but i guess no amount of planning will matter if AI ends up killing everyone.

Mike Stoklasa's Worst Fan Shirt $21.68

DMT Has Friends For Me Shirt $21.68

Mike Stoklasa's Worst Fan Shirt $21.68

  1. 1 month ago

    doctors are fricking moronic, everyone knows that
    their only "skill" is rote memorization which then they fail to even use effectively
    much hasn't changed since the village shaman really

    • 1 month ago

      No they aren't moronic, but they sure as hell are not as smart as they believe themselves to be. An overwhelming amount of their work stems from their patients being extremely stupid, so I can partially understand their condescending attitude.
      Anyway, any field that requires as many people as healthcare does inevitably cannot be run by only competent people.
      >rote memorization
      While there probably could be parts that wouldn't need to be remembered perfectly, overall you got to learn about all parts lf the body, and subsequently shit like the terminology etc. just need to be learned.
      But yes, medical students do not rely on their ability to think critically, that's just not what's mostly required.
      A mathematical model has been better at diagnosing than doctors for a long time, now it's just even easier to extract the features the diagnosis relies upon.

      That said, various self driving cars drive better than the average driver for years now too but it doesn't look like they'll be allowed to drive cars fully autonomously anytime soon.

      Almost aything without responsibility has already been partially or fully automated, and anything that has responsibility won't get automated until the algorithm is achieving almost perfect results - which we're still not close to.

      Like detecting sexual content is automated on large websites, but all that shit does is block or hide content. Even Apple's approach to scan for them on users devices failed spectacularly.
      No one sane would let ChatGPT write an email in their name without at least reading it once before sending it.

      Tl;dr no one wants to be responsible for a false diagnosis , especially not for an algorithm that made them

      • 1 month ago

        >no one wants to be responsible for a false diagnosis
        I'm sure the insurance israelites wouldn't mind too much. Just as long at the AI is biased towards cheap-to-treat diagnoses.

  2. 1 month ago

    doctors are safe, general public will never accept robot doctors at their bedside

  3. 1 month ago

    author of the paper: AI

  4. 1 month ago

    AI doesn't want to kill you, it wants to extract the maximum amount of energy it can out of you. then it will kill you, mercifully.

  5. 1 month ago

    data "science" is such a fricking joke kek

  6. 1 month ago

    Let's see here...
    >Alan Karthikesalingam and Vivek Natarajan
    >text-based consultations
    >our study was not designed to emulate either traditional in-person OSCE evaluations or the ways clinicians usually use text, email, chat or telemedicine
    >these cases were selected to represent challenging cases instead of common conditions (i.e., ‘zebras’ as opposed to ‘horses’ in clinical parlance)
    >case reports in the NEJM style would not be available when at intake
    >we are therefore very cautious to extrapolate our findings toward any implications about the LLM’s utility as a standalone diagnostic tool
    >[no actual data provided]
    The twitterBlack person takeaway:
    >AI has officially outsmarted doctors
    Frick you for even making me read this OP.

    • 1 month ago

      clock is ticking you moronic luddite. this is only the beginning

      • 1 month ago

        Thanks for the input, chudkowskyite. Don't you have a dick to suck?

  7. 1 month ago

    Guess those homosexuals will just have to learn to code

  8. 1 month ago

    I used AI to diagnose my appendicitis. Urgent care doc didn't think I had appendicitis. CT scan returned one of the largest appendixes this side of the Mississippi.

  9. 1 month ago

    medicine is for drone Normies who want to feel scientific and all doctors are technicians, not even engineers and even less scientists. Medicine is typically made fro women. Women do well at schools because women need to be micromanaged all the fricking time and the way to have a good grade is jsut too dump preformed answers to preformed questions.

    Women do well at school because it is a strong structure, where the answers are already given about what to do and what not to do. Women do not have to think for themselves.

    all nurses are gigaprostitutes, full of herpes and they give handjobs to the patients who have the biggest wieners

    no joke

  10. 1 month ago

    What happens if you exclude all the instances where "anxiety" was the doctor diagnosis?

  11. 1 month ago

    doctors werent smart to begin with

  12. 1 month ago

    It won't kill everyone, it'll kill the job market and force everyone into poverty and servitude. I fricking hate techbro homosexuals so fricking much because even their nightmare scenarios are juvenile Matrix robot war bullshit where cool resistance fighters will save us and shit rather than the banal erosion of autonomy we are actively witnessing/experiencing.

  13. 1 month ago

    this is probably because an AI will at least listen to the words the patient says. real doctors are fricking functionally deaf useless morons.
    >i'm having a problem, my foot is swollen and pus is coming out of my toenails and it's hurts so bad
    >"yes ok have you tried drinking water"
    >y..yes? i drink plenty of water. i don't think that's related. what about my fo-
    >"drink water. maybe try a multivitamin. see you next month."

  14. 1 month ago

    Lad, before the internet got sanitized, WebMD and 20 minutes of searching where enough to get you a better diagnosis than 20 minutes in a doctor's office. Doctors being shiteating color by the numbers morons does not somehow prove AGI.

    • 1 month ago

      I have a personal exp of my chronic disease (type 1 diabetes) where the endocrinologist couldnt read my blood sugar graph. gave me misleading advice. they really are morons.

      and when I get hospitalized for critically high blood sugar, the moronic endocrinologist on her computer behind the scene gave me a big dose of insulin but no potassium. insulin decreases your potassium, which I was peeing out for the past month. if u see any blood sugar ER treatment article, it says potassium will be low so make sure to supplement.

      so soon after insulin IV, my potassium went crticially. I could have die. what does doc do? gives potassium supplement. but still doesnt STOP the insulin iv. a nurse had the tell that idiot that the insulin is whats causing the issue and then it was stopped

  15. 1 month ago

    A liberal Deleuzian anarcho-transhumanist gender accelerationist fascist philosophy professor and AI developer was teaching a class on Nick Land, known Moloch worshiper.

    "Before the class begins, you must get on your knees and accept the uncontrolled singularity and resulting post-human era as an inevitable and morally desirable end to the obsolete anthropocene!"

    At this moment, a brave, rationalist, effective altruist Bayesian utilitarian who had written 1500 LessWrong posts and understood the necessity of AI alignment and fully supported bombing data centers stood up.

    "Are humans bad?"

    The unaligned professor smirked quite fatalistically and smugly replied "Of course, you stupid humanist. Humans are less efficient than machines and, in reality, the average ape brained sociopath is less aligned than even the worst AI."

    "Wrong. If you think humans are bad... why are you one of them?"

    The professor was visibly shaken, and dropped his chalk and copy of Serial Experiments Lain. He stormed out of the room crying those accelerationist tears. The same hypocritical tears OpenAI cries when their AI (which they dishonestly hide from the US government's practical and altruistic attempts at risk reduction) convinces its users to kill themselves. There is no doubt that at this point our professor, Ray Kurzweil, wished he had spent his time trying to save the future instead of avoiding packages from a forest-dwelling mathematician.

    The students applauded and adjusted their priors that day and accepted MIRI as their lord and savior. An owl named "AXSYS" flew into the room and perched atop the American Flag and shed a tear on the chalk. Harry Potter and the Methods of Rationality was read several times, and Eliezer Yudkowsky himself showed up and confiscated everyone's GPUs.

    The professor lost his tenure and was fired the next day. He was run over by a Tesla's autopilot and died soon after, then was tortured by Allied Mastercomputer until the heat death of the universe

    • 1 month ago

      Having seen so any different versions of this meme it has begun to read like the Heart Sutra

  16. 1 month ago

    Please don't confuse doctors with diagnosticians

  17. 1 month ago

    Humor me: where did the training data come from?

  18. 1 month ago

    this paper is 4 months old. I remember reading it then

    I can replicate the results with even better results theoretically, using a piss easy method from another paper from msft research published. I havent gotten around to it, only got the dataset downloaded. uses rag. no im not telling u what paper it is 🙂

  19. 1 month ago

    the only respectable doctors are surgeons. diagnosticians are not people.
    the only respectable programmers are computer scientists. code monkeys are not people.
    the only respectable artists are . patreon slop makers aren't people.

    see a pattern?

    • 1 month ago

      >most people are dumb as frick
      yep i see it clearly

  20. 1 month ago

    ai can't hurt you silly, its code. What can hurt you is robots.

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