AI Safety at EAGlobal2017 Conference

Поделиться
HTML-код
  • Опубликовано: 7 сен 2024
  • I attended a charity conference to learn about AI Safety!
    Correction: Alan Dafoe is funded by a grant from the Open Philanthropy Project, but does not work for them.
    The conference's RUclips channel: / @effectivealtruismvideos
    The Website: www.eaglobal.o...
    Jobs at FHI: www.fhi.ox.ac....
    My Concrete Problems in AI Safety series: • Avoiding Negative Side...
    With thanks to my Patrons! ( / robertskmiles )
    Steef
    Sara Tjäder
    Jason Strack
    Chad Jones
    Stefan Skiles
    Katie Byrne
    Ziyang Liu
    Jordan Medina
    Kyle Scott
    Jason Hise
    Heavy Empty
    James McCuen
    Richárd Nagyfi
    Ammar Mousali
    Scott Zockoll
    Charles Miller
    Joshua Richardson
    Jonatan R
    Michael Greve
    robertvanduursen
    The Guru Of Vision
    Fabrizio Pisani
    Alexander Hartvig Nielsen
    Volodymyr
    David Tjäder
    Paul Mason
    Ben Scanlon
    Julius Brash
    Mike Bird
    Taylor Winning
    Ville Ahlgren
    Johannes David
    Andrew Pearce
    Gunnar Guðvarðarson
    Shevis Johnson
    Erik de Bruijn
    Robin Green
    Roman Nekhoroshev
    Peggy Youell
    Konsta
    William Hendley
    Adam Dodd
    DGJono
    Matthias Meger
    Scott Stevens
    Michael Ore
    Robert Bridges
    Dmitri Afanasjev
    Brian Sandberg
    Einar Ueland
    Lo Rez
    Stephen Paul
    Marcel Ward
    Andrew Weir
    Pontus Carlsson
    Taylor Smith
    Ben Archer
    Ivan Pochesnev
    Scott McCarthy
    Kabs Kabs Kabs
    Phil
    Christopher Askin
    Tendayi Mawushe
    Gabriel Behm
    Anne Kohlbrenner
    Jake Fish
    David Rasmussen
    Filip
    Bjorn Nyblad
    Stefan Laurie
    Tom O'Connor
    pmilian
    Jussi Männistö
    Cameron Kinsel
    Matanya Loewenthal
    Wr4thon
    Dave Tapley
    Archy de Berker
    / robertskmiles

Комментарии • 105

  • @lebagelboy
    @lebagelboy 6 лет назад +18

    loving the subtle fray song at the end - very fitting.

    • @nowheremap
      @nowheremap 5 лет назад +1

      How to save a life... or many, or humanity itself, and as cost-effectively as possible

  • @morkovija
    @morkovija 6 лет назад +65

    Cool shoutout! Someone is famous!

    • @MegaHarko
      @MegaHarko 6 лет назад +5

      jesus... At first I read: School shootout! Someone is famous!
      what's wrong with me? ⊙﹏⊙

    • @zamundaaa776
      @zamundaaa776 6 лет назад +1

      Hans Wurst you've been too long in school?
      This illness can be fixed by playing games a recommended time of minimum 2 hrs

  • @Nulono
    @Nulono 4 года назад +5

    I love how one of the pros at 2:40 is "good fluid dynamics".

  • @TheSpacecraftX
    @TheSpacecraftX 6 лет назад +35

    Everything is terrible
    - 7.8/10 IGN

    • @mrsuperguy2073
      @mrsuperguy2073 6 лет назад

      TheSpacecraftX every mayor's terrible!

  • @alisdairhodgson9905
    @alisdairhodgson9905 6 лет назад +12

    Totaly hooked on your videos.

  • @Coly0101
    @Coly0101 6 лет назад +3

    Good job on getting the shout out! Interesting ideas. Looking forward to seeing the videos go live.

  • @666lumberjack
    @666lumberjack 6 лет назад +40

    I loled @ 7.8/10

  • @ToriKo_
    @ToriKo_ 6 лет назад +3

    BostonDynamic and Robert Miles upload on the same day? Amazing!

  • @arponsway5413
    @arponsway5413 6 лет назад +3

    another nice video mate. keep up the good work. love from Bangladesh.

  • @TribeWars1
    @TribeWars1 6 лет назад +1

    From the title i thought this would be a boring video, but it was actually really great.

    • @brewbrewbrewthedeck4138
      @brewbrewbrewthedeck4138 6 лет назад

      I'm kinda jelly that he got to go there. Cashews by the fistful? Count me in!

  • @himselfe
    @himselfe 6 лет назад +12

    Is the 1 dislike from the person that didn't like the cashews?

    • @fleecemaster
      @fleecemaster 6 лет назад +3

      1 dislike from the person that didn't get to eat the cashews because Rob got to them all first

    • @JM-us3fr
      @JM-us3fr 6 лет назад +3

      The other 2 were from an AI

  • @propsz
    @propsz 6 лет назад +6

    That IGN reference

  • @tombstone909
    @tombstone909 4 года назад +2

    2:40 "It has a little something for everyone"

  • @hypersapien
    @hypersapien 6 лет назад

    Thank you Kyle Scott!

  • @PaceMethod
    @PaceMethod 6 лет назад

    Nitpick: Alan Dafoe is from Yale and works with the Future of Humanity Institute. While he is funded by a grant from the Open Philanthropy Project, he does not work *for* them.

    • @RobertMilesAI
      @RobertMilesAI  6 лет назад +1

      Thanks for the correction Ben, I've added it to the video description :)
      I don't know where I got that from actually, it's all pretty clear on his title slide. Embarrassing.

  • @ZachAgape
    @ZachAgape 4 года назад +1

    "and he's right about everything" *I'd strongly recommend u to watch Robert Miles* xD
    good vid, thanks for referring to Effective Altruism!

  • @cornjulio4033
    @cornjulio4033 11 месяцев назад

    Talk to us Robert, we miss you !

  • @TylerMatthewHarris
    @TylerMatthewHarris 6 лет назад +3

    If strong AI comes to be a thing there is a small chance that it would arise from something like the mandelbrot set formula. If this were the case it would be possible for an individual to stumble upon the answer. You should make a "if you find the answer to general AI" video. Because there are no videos like it, and I'm sure they will be wondering why they should/shouldn't run the program.

    • @the1exnay
      @the1exnay 4 года назад

      That is an interesting idea. To find the solution to general intelligence while coding something seemingly unrelated
      It's probably good that I'm unlikely to stumble upon anything like that since my approach to programming is "the best way to figure out what I've done wrong is to run it and see what happens"
      Now I'm just imagining someone finishing some complex code to figure out P vs NP or whatever and after running for a few minutes a pop up appears giving some excuse for why the computer needs internet access in order to resume calculating and then we're doomed.

  • @DamianReloaded
    @DamianReloaded 6 лет назад +9

    Cashew is the banana of nuts.

    • @jon2431
      @jon2431 6 лет назад +4

      Cashew for scale?

    • @dariusduesentrieb
      @dariusduesentrieb 6 лет назад

      i dont like cashews and i dont like banans

    • @Kram1032
      @Kram1032 6 лет назад

      I don't care much for cashews but I love bananas.

    • @riplememe4460
      @riplememe4460 6 лет назад +1

      Because they look alike? For real though banana is the potato of fruit from a culinary perspective and not enough people say it

    • @litarea
      @litarea 6 лет назад

      Except cashews are really expensive, like $12 a pound

  • @ausbare140
    @ausbare140 4 года назад

    How will we stop big companies from using AI to increase profits when the bottom line is all that big companies think about.

  • @windar2390
    @windar2390 6 лет назад +2

    1:52 - "Talk to 80k people in person". excuse me, wtf?

  • @General12th
    @General12th 6 лет назад +3

    *Where did I go wrong?*
    *I LOST A FRIEND*

    • @fleecemaster
      @fleecemaster 6 лет назад

      They are always in the last place you look

    • @AliceDiableaux
      @AliceDiableaux 6 лет назад

      Every time I hear any version of that song I think about that Scrubs episode and tear up :')

  • @michalbreznicky7460
    @michalbreznicky7460 5 лет назад

    I was a bit surprised that these guys consider AI safety a higher-priority problem than AGW. I don't necessarily disagree, though I would like to see the reasoning behind this.

  • @tomahzo
    @tomahzo 2 года назад

    "... Not REALLY!" ;D

  • @PaperBenni
    @PaperBenni 5 лет назад +3

    Effective Altruism has an unfortunate acronym.

  • @CutcliffePaul
    @CutcliffePaul 4 года назад +1

    1. The worst VR experience of your life? Hopefully, you mean because the VR experience was terrible because it worked really well and showed you just what it's like to be a factory-farmed animal, not because their VR implementation was the worst you've seen, right?
    2. The Future of Humanity Institute? Wow, that's a pretty important-sounding organisation, right? 🤯

  • @jordanfry2899
    @jordanfry2899 6 лет назад +2

    I highly recommend going vegan. It's a remarkably cost-effective way effort-wise to make the largest positive impact on the world. Animal agriculture is one of the leading causes of global warming, not to mention the direct environmental impact it has such as polluting rivers and unprecedented levels of deforestation. It's one of the best changes I've made to my life.

    • @mvmlego1212
      @mvmlego1212 5 лет назад

      Animal agriculture is only responsible for 9% of greenhouse gas emissions, compared to 28% from electricity and 29% from transportation. I agree that its other environmental impacts are worse than those of fossil fuel industries, though.

    • @howtoappearincompletely9739
      @howtoappearincompletely9739 Год назад

      I tried and couldn't hack it, unfortunately. I hope to see lab-grown meat become a big thing in the near future.

  • @thefridgefreezer
    @thefridgefreezer 6 лет назад

    Owine Evans. And do you? No I don't.

  • @julianw7097
    @julianw7097 6 лет назад +14

    iAnimal turned me vegan. Definitely worst VR experience ever.

    • @oktw6969
      @oktw6969 6 лет назад +3

      Do you know plants have a rudimentary nervous system and even "communicate" between each other? In fact the "smell" of freshly cut grass is comparable to smell of blood in slaughterhouse. And if you plan to eat rocks instead, they have a lot of microorganisms on the surface, which most likely won't survive through the gastric acid (the slaughterhouse in your body basically).
      But a shitty VR game turned you vegan indeed.

    • @julianw7097
      @julianw7097 6 лет назад +5

      +Oktavia Von Seckendorff I have heard a bit about claimed plant subjective experience, however I haven't read any of the science on the topic (didn't find any). Are you aware of any research? If you have any I am happy to review it and change my actions accordingly.
      It should also be noted that the VR experience was simply the first step. To fully change my moral philosophy took many days.

    • @lenn939
      @lenn939 6 лет назад +10

      Von Seckendorff That is complete esoteric bullshit. Plants don’t have a CNS, and I have no idea where you would get that from. “Communication” and sensory perception don’t imply sentience whatsoever. Computers can communicate with each other and we have created peripherals through which they can sense inputs. That’s how it works with plants.
      I could destroy your argument even if it was true that plants are just as sentient as animals. What do you think it is that we feed our cattle? Energy gets lost as you go up the food chain, so you’d need to kill more plants to get the same amount of energy from meat than just directly eating the plants. Even then it would minimize suffering to abstain from meat consumption.
      Vegans aren’t concerned about microorganisms so I’m not sure why you bring them up. It’s not about going to extremes to never kill any living being ever. Vegans care about minimizing suffering, that’s all. If humans really required animal flesh to survive then vegans would be about treating our cattle as humanely as possible. We don’t need animal flesh though so we can do even better and just abstain from meat eating.

    • @lenn939
      @lenn939 6 лет назад +10

      Oktavia Von Seckendorff You really are not the brightest tool in the shed, are you? Vegans don’t care about whether a thing is alive or not alive, vegans care about sentience, or the ability for an organism to suffer. As far as we know, that requires a nervous system, something which plants lack as much as our current computers. There’s no reason to assume they have any ability to suffer, it wouldn’t even make sense from an evolutionary standpoint.
      Contrary to your claim, this channel has never discussed the emergence of sentience in AI. So far this channel has only focused on the topic of intelligence, which doesn’t necessitate sentience.
      The project of minimizing suffering is not as subjective as you make it out to be. We do have an idea what gives organisms the ability to suffer, this can be studied to some extent. If we find out anything new then that is going to be incorporated into the vegan project.
      Again, we have good reason to assume that cows can experience much greater suffering than dust mites, so it makes sense that we are more concerned about them. And like I said, veganism isn’t about going through extreme lengths to save every animal on the planet. We all agree that there are limitations, nature is cruel and some animals are going to suffer regardless of what we do. We can’t look after every dust mite. Veganism is grounded in reality, it’s about doing what is practically possible to minimize suffering. What you are proposing is called the nirvana fallacy (www.logicallyfallacious.com/tools/lp/Bo/LogicalFallacies/134/Nirvana-Fallacy ). Just because we can never achieve the ideal doesn’t mean that it isn’t better to try to do as much as is practically possible.

    • @nickappleton3968
      @nickappleton3968 6 лет назад +11

      Oktavia Von Seckendorff so the logical default is veganism. It takes 6 pounds of plant matter to produce one pound of animal matter. So if you're concerned about the welfare of plants, go vegan and you'll kill one sixth the number of plants.

  • @pacibrzank78
    @pacibrzank78 6 лет назад +1

    Concrete Problems in AI safety Part 3.5 minute 1:50. just saying.

    • @HansPeter-qg2vc
      @HansPeter-qg2vc 6 лет назад +1

      I don't understand what you want to say with this. What part of this video are you attempting to apply that reference to?

  • @KingOfChaos213
    @KingOfChaos213 6 лет назад

    Not really...

  • @derstreber2
    @derstreber2 6 лет назад +1

    I don't know. Safety with regards to building anything is an important thing to consider, but once you start making legislation and policies that dictate what I can and cannot program in the privacy of my own home, that is where I must voice concern. For the same reason I voice concern over DRM, software patents, and the actions taken by software companies to take power away from the user, I don't think the government, an organisation, or a body of organisations should be the ones deciding who gets a permit to do research about AI or research about anything. If AI is so important and powerful, why would we want to give that power to an exclusive group. The spirit of freedom, openness, and scientific inquiry seem diametrically oppose to such laws.
    Corporations and individuals will naturally want to make safe decisions. If they don't, we have other legal means of correction once someone gets hurt or damages public or private property. There is already a lot of incentive not to act haphazardly.

    • @howtoappearincompletely9739
      @howtoappearincompletely9739 Год назад

      See Robert Miles's "The other 'Killer Robot Arms Race' Elon Musk should worry about": ruclips.net/video/7FCEiCnHcbo/видео.html for an explanation of why the approach you advocate is dangerously naïve.

  • @dport
    @dport 6 лет назад +10

    Was it enough to turn you vegan? 🌱 Or are you already? 😯

    • @badkitty3559
      @badkitty3559 6 лет назад +9

      My steak tastes great

    • @AliceDiableaux
      @AliceDiableaux 6 лет назад +12

      Bad Kitty, Oh yeah, taste buds are so much more important than the suffering of sentient animals on such a gigantic scale that it isn't really compareable to anything else. Nice going, so much empathy, not. News flash: no one above the age of 15 thinks it's cool to be a sociopath.

    • @dport
      @dport 6 лет назад +7

      Bad Kitty Ugh, what an over done response. Saying that I don’t want to start a flame war, I just want to know if he’s vegan :)

    • @DirtyPoul
      @DirtyPoul 6 лет назад +1

      Daniel Porteous I agree with you. AliceDiableaux' took it to a whole other level though.

    • @brewbrewbrewthedeck4138
      @brewbrewbrewthedeck4138 6 лет назад

      Prove to me the sentience of animals and then we'll talk. All there is right now are mere suggestions based on spurious similarities to humans, tainted by an unhealthy dose of anthropomorphization. For all we know the internal emotional depth of animal might be closer to that of a rock than that of a human. Giving off the appearance of experiencing suffering is not the same as actually feeling.
      I mean dipshit "animal rights activists" would have you believe that lobsters being cooked alive are screaming (lol, with what vocal chords?) when the sound is just water evaporating. Similarly, you cannot show the presence of qualia and sentience just by observing an animal’s actions. That’s behaviorism and I don’t think vegans wanna go back to that.
      In the meantime, I’ll save my empathy for beings that I actually know deserve it, namely my fellow humans. Somehow quite a lot of vegans are utterly and reprehensibly failing on that front. But I guess this is just a natural consequence of adopting the retarded “all life is equal” stance and actually caring about animals just as much (if not more) than about their fellow men.

  • @gummansgubbe6225
    @gummansgubbe6225 6 лет назад +2

    Pøhh. These people know what they are doing. Just look at their diner expenses. 10000 dollar for a dinner for two. Dont give them money. Look at their expenses!