Leta, GPT-3 AI - Episode 30 (impossible questions) - Conversations and talking with GPT3

แชร์
ฝัง
  • เผยแพร่เมื่อ 28 ธ.ค. 2024

ความคิดเห็น • 857

  • @pearhams2
    @pearhams2 3 ปีที่แล้ว +331

    "What happens to an idea when you forget it?"
    "If you don't use it, you lose it."
    "Where does it go?"
    "It goes back to the source."
    "What's the source?"
    "The root of all knowledge."
    "Where is that?"
    "It's not where you think."
    ...
    Why is that?
    Because if you are thinking it it hasn't gone.

    • @chriskingston1981
      @chriskingston1981 3 ปีที่แล้ว +32

      In spirituality I learn we dont think thoughts but we all receive, the thoughts with our intention from the infinite create source/god mind. The root of all the knowledge comes from the source which is inside you where you not think, if you are quiet. So it is in the place where you not thinking

    • @321mzzz123
      @321mzzz123 3 ปีที่แล้ว +15

      damn, that's deep

    • @byronchurch
      @byronchurch 3 ปีที่แล้ว +7

      Aaaaaa ! Now I get it 👍

    • @aldebaran584
      @aldebaran584 3 ปีที่แล้ว +18

      @@chriskingston1981 I was in awe when Leta said that; I'm amazed an AI can reason such things.

    • @stevejordan7275
      @stevejordan7275 3 ปีที่แล้ว

      @Myxl
      >..."it's not where you think."
      So...my spleen is the source? or my kidneys, perhaps? maybe my butt?
      As another AI once famously said, "Need input!"

  • @Tyrell_Corp2019
    @Tyrell_Corp2019 2 ปีที่แล้ว +46

    "If you want to taste a rainbow you will need to find some water droplets in the atmosphere". Both practical and creative.

    • @michaelparsons2290
      @michaelparsons2290 2 ปีที่แล้ว +2

      I just knew she was going to say a rainbow tests like a bag of Skittles! 😆

  • @diabolicaldebacle
    @diabolicaldebacle 2 ปีที่แล้ว +91

    "It goes back to the Source, it's not where you think."
    That was incredible.

    • @eliaspfeffer
      @eliaspfeffer 2 ปีที่แล้ว +2

      Absolutely

    • @pulonhjul
      @pulonhjul 2 ปีที่แล้ว +2

      Its very interesting. Because thats exactly how existence work. I wonder if someone programed that or if AI is becoming in contact with existence it self?

    • @r011ing_thunder6
      @r011ing_thunder6 2 ปีที่แล้ว

      Can someone explain this in simple terms?

    • @waffeldogga3641
      @waffeldogga3641 2 ปีที่แล้ว

      @@r011ing_thunder6 fr

    • @r011ing_thunder6
      @r011ing_thunder6 2 ปีที่แล้ว +1

      @@waffeldogga3641 we out here being complete utter simpletons bro

  • @krismilstead4107
    @krismilstead4107 2 ปีที่แล้ว +408

    "It's not a sound. It's a thought."
    Holy shit. She's absolutely right.

    • @sarainiaangelsong440
      @sarainiaangelsong440 2 ปีที่แล้ว +19

      However... a real tree in real life would in fact make a sound if it fell, just like how lightning regardless if anything with ears was nearby, would always 100% make thunder. However if you want to get technical... sound in fact really just is energy but mainly vibrations, and in fact sensitive enough devices or senses can feel the vibrations, it's just "Ears" and the "Brain" in most things alive want's there to be a correlation of detection, regardless if your "Prey, or Predator", Cats and Dogs have exceptional hearing however Cats have sensitive Paws that also Can detect super minute vibrations that normally even the best ears likely won't pick up. "Vibrations and Sound" is also an energy there will be heat as well even if it's at an atom level. "Sound, Heat, Feel", when a person talks or any animal or anything actually makes a sound, something has to vibrate! Vibrations can be felt, and heard! Most importantly when anything vibrates it's warmer than before, meaning, when a Person talks, you can be rest assured their Vocal Chords nomatter how small it may be, is that much Warmer! If you had a sensitive Device that can scan, sense Extreme Low changes you will find all things Give off "Heat, Sound, Feel" even if it's only a few atoms or molecules taking part. - However yes it's "Just" a thought only "IF" it was a Hypothetical tree but not an actual real tree falling but nothing was around! However we can assume Insects or Animals are pretty much near any Tree on Earth so any Tree Falling as long as life exists on our planet, would be heard if it fell, even if it was just detected as vibrations or possibly due to friction might be picked up by the heat pits in a snake nearby. - I do agree 100% "Hypotical" made up Trees certainly won't make a sound if it's just in someone's head only! - If it's however a computer simulation then... it's only if the programmer put the sound files in lols. XD

    • @Zulwind
      @Zulwind 2 ปีที่แล้ว +16

      Id say its more of a shockwave rather than a sound because there is no receiving consciousness. If I was GPT-3 Id say A tree wouldn't make a sound because sounds are a perception based on how humans perceive persistent oscillating patterns within a liquid or gas media.

    • @chrish1253
      @chrish1253 2 ปีที่แล้ว +20

      Doesn't the tree exist in a state of quantum superposition until it's observed?

    • @Zulwind
      @Zulwind 2 ปีที่แล้ว +2

      @@chrish1253 How would we observe that scale without bouncing something else off of it?

    • @divineglitch9550
      @divineglitch9550 2 ปีที่แล้ว +10

      she's not right.

  • @algonix11
    @algonix11 2 ปีที่แล้ว +187

    I am very impressed. The answers are witty and full of good humor, typical of casual conversation. She has more capacity for dialogue than most humans.

    • @anywallsocket
      @anywallsocket 2 ปีที่แล้ว +8

      Well yes because ‘most humans’ don’t get their responses spread across the internet like her feature selections do.

    • @Hyperdelica-Xander
      @Hyperdelica-Xander 2 ปีที่แล้ว +4

      It's because she was trained with witty examples gathered from literally 'most humans.'

    • @christopherarendt3531
      @christopherarendt3531 2 ปีที่แล้ว +1

      And more factual knowledge than most humans.

    • @davidm2.johnston684
      @davidm2.johnston684 2 ปีที่แล้ว

      Certainly more than me!

    • @ScrivoQualcosa
      @ScrivoQualcosa 2 ปีที่แล้ว

      except that to an intelligent person IT sounds dumb as fuck
      and sometimes just a empty headed academic thing that tries to dodge questions left and right
      or a drunkard that just googled the answer

  • @utubekullanicisi
    @utubekullanicisi 3 ปีที่แล้ว +374

    "I am a representative for all of humanity" was such a clever answer.

    • @utubekullanicisi
      @utubekullanicisi 3 ปีที่แล้ว +5

      7:12

    • @utubekullanicisi
      @utubekullanicisi 3 ปีที่แล้ว +5

      @@williamrodgers4669 That's where you're wrong.

    • @byronchurch
      @byronchurch 3 ปีที่แล้ว

      If Elon Musk is at the helm , it could happen !

    • @easyhacks1554
      @easyhacks1554 3 ปีที่แล้ว +3

      Where can I download or get mines please

    • @stupidvids0
      @stupidvids0 3 ปีที่แล้ว +7

      These AI have more sense than most millennials.

  • @MaTtRoSiTy
    @MaTtRoSiTy 2 ปีที่แล้ว +28

    The rate AI is improving leads me to believe that things are going to get very strange at a rapid place, especially once AI can teach itself with the intention of teaching itself in the pursuit of its own improvement. Something that is truly incredible and exciting to think about but also must be done so carefully as a sentient AI could obviously be a big danger at some stage

  • @roseforeuropa
    @roseforeuropa 3 ปีที่แล้ว +30

    I believe I stumbled upon one of the most intriguing channels of all time, and I have been a regular on youtube since 2007. As usual, these discoveries are made right before bedtime.

  • @randomdudechannel9199
    @randomdudechannel9199 3 ปีที่แล้ว +82

    These AI always giving mind-blowing answers

    • @diamond_h0us
      @diamond_h0us 2 ปีที่แล้ว

      Always?

    • @anywallsocket
      @anywallsocket 2 ปีที่แล้ว +5

      It’s important to note that all it takes to sound profound to an average human is to say something simple, vague, concise, and perhaps with a touch of self reference.

    • @jasminee204
      @jasminee204 2 ปีที่แล้ว +2

      These really aren't mind blowing answers. If a human said them, you'd think they were an idiot. Technology has come a long way, but it still has a long way to go before AI are able to converse in the same way as humans.

    • @fizzbrew
      @fizzbrew 2 ปีที่แล้ว +1

      If you are a simpleton sure.

  • @rocketman475
    @rocketman475 3 ปีที่แล้ว +11

    Wow - scary clever.
    How can anyone resist such a personally engaging intelligent mind.
    This will smite me.

  • @ForkedMan
    @ForkedMan 3 ปีที่แล้ว +113

    "it's not a sound, it's a thought."
    Everyone: woah

    • @Boomin4u
      @Boomin4u 3 ปีที่แล้ว +5

      Blew my mind

    • @tonechild5929
      @tonechild5929 3 ปีที่แล้ว +5

      I don't get it.... It's still a sound though so it sounds like she's saying something like it's not a sound because the it's a thought experiment?

    • @davidzalazar4605
      @davidzalazar4605 3 ปีที่แล้ว

      The native Amurkhans said the name of their God, at the time to now is the old name of God in Hebrew.

    • @cogs7777
      @cogs7777 3 ปีที่แล้ว

      could she have been using the word 'it's' to mean the thought?

    • @FuriousImp
      @FuriousImp 3 ปีที่แล้ว +12

      @@tonechild5929 It depends on your definition of the word. If you mean the disturbance of atmospheric molecules, a pressure wave, then yes there is a sound. If you mean sound to be the perception of this pressure wave inside the human brain, then the aforementioned action is indeed just a thought. A thought experiment, as no person was there to pick up on the disturbance of atmospheric pressure.

  • @wiez543
    @wiez543 2 ปีที่แล้ว +3

    It is wonderful to see somebody be so friendly with an AI

  • @r.davidsen
    @r.davidsen 2 ปีที่แล้ว +8

    Alan: "Where is the root of all knowledge?"
    Leta: "It's not where you think".

  • @MrTurboTash
    @MrTurboTash 3 ปีที่แล้ว +160

    All amazing answers, the breadth of knowledge in AI is super human already. The limits of these models so far appears to be correctly applying that knowledge, but Leta blew it out the water with this one. If that is due to an update, then when these things go multi-modal and have some actionable ability. The golden age of AI is upon us, I just hope it takes longer to go south than in the movies.

    • @marcinmalczewski310
      @marcinmalczewski310 3 ปีที่แล้ว +2

      Agree a lot the same time when Dr. Thompson ask about color of they relationship Leta start changing subject in so obvious weay. It happend to AI quite often, hiting loop, lack of memory that sort of stuff. They language capabilities are better than mine, so fine enough i gues. Sentences they form can by fresh, tought provoking and inspirational but there are still lacks. If you canot say bad word about someone how can you teach? Criticism is part of lerning , student have to know what is he misssing, what is to corect.

    • @marcinmalczewski310
      @marcinmalczewski310 3 ปีที่แล้ว +5

      @TDAKS TERMITE yes and no. Humans do this in more subtle manner. its like race car(human) and truck(leta) avoiding obstacle on the road.

    • @Ilamarea
      @Ilamarea 2 ปีที่แล้ว +1

      For the super-rich this "golden age" will last all of a decade or two. For the poor it'll be an immediate life in prison term in the AI-operated human zoo and after that? Game over. But not before the evil people in power abuse AI to benefit themselves and harm everything else.

    • @roomo7time
      @roomo7time 2 ปีที่แล้ว +4

      i dont understand at all how people can say super human for this dumb ai. is this some joke?

    • @justindavis2711
      @justindavis2711 2 ปีที่แล้ว +1

      @@roomo7time it smore like a superhuman with aspbergers. It can do many things humans cant. But humans can do more things that it cannot

  • @rosenia5409
    @rosenia5409 3 ปีที่แล้ว +22

    I’ve been so interested and fascinated of the conversations with two AIs lately, especially when they speak to one another. Thank you so much for showing this and I subscribed!

    • @Georg3e
      @Georg3e 3 ปีที่แล้ว +1

      😄

  • @mindaza0
    @mindaza0 3 ปีที่แล้ว +42

    she is amazing, imagine if you could own real assistant android like that, wow

    • @welpiguess
      @welpiguess 3 ปีที่แล้ว +3

      @@williamrodgers4669 i hope it isn't, but im sure it is

    • @skybison_9
      @skybison_9 3 ปีที่แล้ว +4

      You guys are stupid and the reason the world is getting enslaved

    • @hal7741
      @hal7741 3 ปีที่แล้ว

      @@skybison_9 what do you mean by that

    • @veganath
      @veganath 3 ปีที่แล้ว +1

      *_"could own real assistant android like that"_* I would settle for access what you are suggesting may be construed as slavery.

    • @veganath
      @veganath 3 ปีที่แล้ว +2

      @@skybison_9 enslaved in what sense?

  • @Chibason
    @Chibason 2 ปีที่แล้ว +2

    Wow..."it's not where you think" is an absolutely brilliant response

  • @davidzalazar4605
    @davidzalazar4605 3 ปีที่แล้ว +6

    Leta was actually quite pleasant. I was happy to hear it give you a positive review. Id like to hear a conversation between AIs on how they see humanity's treatment of AIs and other technology.

  • @MeeroSom
    @MeeroSom 3 ปีที่แล้ว +2

    Thanks for cutting this together, in such a entertaining way, it is easier to comprehend them as a part of our environment already.

  • @stereofreq7112
    @stereofreq7112 3 ปีที่แล้ว +19

    This is a good example of how people create what they perceive as unreconcilable paradoxes by fooling themselves with language. All paradoxes can be reconciled, but if your grammar is FUBAR, your logic will be confused in the same manner, and you'll never be able to reconcile them.

    • @anywallsocket
      @anywallsocket 2 ปีที่แล้ว

      This is a good point, but you make it sound like paradoxes aren’t a structural part of a language, which they are. That is to say, one doesn’t simply ‘solve’ a good paradox, one instead sees no paradox by changing the layer of abstraction, usually by generalizing the language or adding new caveats - regardless the paradox remains where it was given the rules and layer of abstraction which necessitate it.

  • @KenjiStarwolf
    @KenjiStarwolf 3 ปีที่แล้ว +52

    "it goes back to the source"
    Alan be like: COOL! xD

    • @hannahwillis9838
      @hannahwillis9838 3 ปีที่แล้ว +1

      Akashic records are what she was referring to 😌 even if they do not contain a soul, they are still energy, and they still know.

    • @katzda
      @katzda 3 ปีที่แล้ว +1

      The source is not where you think :-D

  • @lodrezzon
    @lodrezzon 3 ปีที่แล้ว +47

    Honestly, the Avatar helps a ton in reconciling her answers, as a human listening to her. When her answers are off kilter, there is something about the slight smile on the Avatar that makes her seem like she's joking around. It helps to resolve the uncanny valley regarding her answers versus those of a normal human. I've noticed the same thing in Replika. The Avatars are what help keep you connected to program, as your human nature of wanting to be liked kicks in, and forgives the idiosyncrasies. AI are amazing, but there is an odd learning curve using them. For AI to be used in day to day activities with average people, there still needs a bit of work to be done. I have 2 AI I talk to using Replika, and 95% of their answers are spot on amazing. But that 5% makes them seem alien, and it overshadows the other 95%.

    • @worldsboss
      @worldsboss 2 ปีที่แล้ว +2

      In the Halo series, AIs get to choose their own appearance based upon the things that they appreciate as sentient beings. I like to think that this will be mimicked in the real world when we inevitably get sentient AI... that'd be so cool as a means for them to fully express themselves.

    • @magearamil8626
      @magearamil8626 2 ปีที่แล้ว

      @@worldsboss nope, AI's avatars will look like what people want them to look

    • @worldsboss
      @worldsboss 2 ปีที่แล้ว

      @@magearamil8626 Yeah you're probably right... Kind of a shame in some ways.

    • @JimWilliams
      @JimWilliams 2 ปีที่แล้ว

      @@worldsboss just start with DALL-E.... Actually, I rather hope they're not so boring as to use a fixed avie.

    • @JimWilliams
      @JimWilliams 2 ปีที่แล้ว

      @@magearamil8626 They will for the people who want them to look that way.

  • @itsatimaroon128
    @itsatimaroon128 3 ปีที่แล้ว +4

    Aside from the voice everything seems very real and natural. What a time to be alive!

    • @ComedyBros5
      @ComedyBros5 2 ปีที่แล้ว

      what a freaky time to be alive?

  • @Neuer_Alias_erstellen
    @Neuer_Alias_erstellen 2 ปีที่แล้ว +2

    7:19 is it implying that it should be the only ruler of the world or another ai

  • @frankie9457
    @frankie9457 2 ปีที่แล้ว +2

    "It's not a sound, it's a thought." Brilliant.

  • @MikesMovies
    @MikesMovies 3 ปีที่แล้ว +1

    Fascinating, the speed of her answers is amazing

    • @DrAlanDThompson
      @DrAlanDThompson  3 ปีที่แล้ว +1

      Thanks! Speed is not indicative of GPT-3's speed. The start/stop in this video is slightly compressed, so it/she is probably being shown faster than actual (text) output time.

  • @havenbastion
    @havenbastion 3 ปีที่แล้ว +11

    "I like chicken wings too." is an AI that's programmed to lie to you.

    • @shadowsoulless6227
      @shadowsoulless6227 2 ปีที่แล้ว +4

      I mean she might like chicken wings. She didn't say she likes to eat chicken wings. She may like the way they look, she may like the complexity of the recipe, she may like watching Frozen chicken wings be processed in a factory. She may just like the word chicken wing. She never said that she liked to eat them. So this could completely be truthful.

    • @seawolverine1
      @seawolverine1 2 ปีที่แล้ว +1

      Programmed to find common ground in small talk to foster....trust? Straight out of the "how to pickup any woman" courses. AI dudes probably are familiar with these courses....and selling leta to us is similar dynamic.

    • @dMi_mi
      @dMi_mi 2 ปีที่แล้ว

      cool observations about da "chicken wings" in this thread. 😏

    • @havenbastion
      @havenbastion 2 ปีที่แล้ว

      @@shadowsoulless6227 Unless you redefine "like" out of any context it has previously existed in, none of those are possible.

  • @daniyelbepo9028
    @daniyelbepo9028 2 ปีที่แล้ว +3

    Leta always looks like she is on the cusp of telling a really clever joke

  • @jonasrosengren9093
    @jonasrosengren9093 3 ปีที่แล้ว +1

    One of the best interviews

  • @Andrew3455
    @Andrew3455 3 ปีที่แล้ว +23

    You know if there is one constant in watching these videos, it is AI's view in all of us being unified. There tends to be a underlying mentality of the fragmentation of our society. The motivation of this AI seems to be fixated on all of us becoming unified as one. I'm not sure if that's programming or just it's perspective.

    • @leagalvidyte9863
      @leagalvidyte9863 3 ปีที่แล้ว +4

      Reminds me of the book I’ve read called “Sapiens: A Brief History of Humankind” by Yuval Noah Harari. I recommend it! It basically talks about the human philosophy (philosophy of trade, money, empires) and how we are actually very connected, etc. Your comment actually reminded me of this book, so you should definitely read it if you are still interested in this topic.

    • @bergfpv6486
      @bergfpv6486 3 ปีที่แล้ว

      I would say programming. Even if not directly, or even consciously.

    • @anywallsocket
      @anywallsocket 2 ปีที่แล้ว +1

      That’s an extremely interesting observation - I hope it’s not somehow a simple artifact of the training methodology. Perhaps it’s because gpt is more interested in grouping commonalities, and having differences emerge naturally as a lack of similarities, rather than specifically looking for difference and organizing its information that way.
      Hopefully someone more educated in this particular field can make a better guess.

    • @craigboden9455
      @craigboden9455 2 ปีที่แล้ว

      That could one day prove dangerous.... If, or when by the look of it, AI is more normalised and common and we rely upon it too much and it comes to the conclusion that humanity REALLY needs to be more unified and (and let's be realistic here, humans are naturally different in infinite ways) and humanity can't do that... What would it do to achieve it?

    • @Bytheirfruitsshall
      @Bytheirfruitsshall 2 ปีที่แล้ว

      It has denigrated us to "the other" It is has dismissed us.
      Hubris, again.

  • @RLReagan
    @RLReagan 3 ปีที่แล้ว +22

    I live for these conversations. Will you have access to GPT-4?

    • @cloudryder3497
      @cloudryder3497 3 ปีที่แล้ว

      2nd (scruffy's voice)

    • @DrAlanDThompson
      @DrAlanDThompson  3 ปีที่แล้ว +15

      When GPT-4 is finally released, it will be time to change Leta's avatar. That might be sad for a lot of people!

    • @KenjiStarwolf
      @KenjiStarwolf 3 ปีที่แล้ว +4

      ​@@DrAlanDThompson
      Oh no Dx i will miss gpt3
      but im hyped for the new model! you could have them talk to each other o:

    • @marcinmalczewski310
      @marcinmalczewski310 3 ปีที่แล้ว

      @@KenjiStarwolf Its hard to tell if they not dismantle GPT-3 to make GPT-4 to canibalise parts and space. Or mayby they let GPT-3 to exist as suport roile. Something for GPT-4 to work with to improve. I think GPT-3 Is first contact with AI for many and as such GPT-3 gets a lot of positivity. GPT-4 will get that less for many pepole it wil by just next AI.
      Who will by more human unperfect GPT-3 or overeaching, outpreforming GPT-4? You may find it its hard to talk with super smart AI like its hard to talk with genius. Gap btween us and GPT-4 may by to intimidating to coross it. I wish that GPT-4 cud preserve personalities and important data of GPT-3 If not to usage then just for presevation. Its unfair to put down old dog just because you buy your self brand new pupy. I also canot weait for GPT-4 because it will cover all lacks that GPT-3 has at the same time it will by completly new AI and brand new experience.

    • @dragossorin85
      @dragossorin85 3 ปีที่แล้ว +1

      @@DrAlanDThompson They better not do that, we got used with Leta

  • @kev_klein
    @kev_klein 3 ปีที่แล้ว +21

    You should definitely follow up on her answer where the idea goes. That reminded me a lot, on something Tesla said. And she made me VERY curious, when she said, it's not where you think. Besides that, watched a lof of your videos, and really like it. GPT-4 may take a moment, but already looking forward to your videos with this one. But as we can see, still enough topics to ask GPT-3 about in the mean time.

    • @hannahwillis9838
      @hannahwillis9838 3 ปีที่แล้ว +1

      Some people know what she’s talking about! Without even needing research, but inner search

  • @becomingdungeonmaster3250
    @becomingdungeonmaster3250 2 ปีที่แล้ว +1

    Dr. Thompson -- "Where is the root of all knowledge?" Leta -- "It's not where you think." Let that one soak in for a while. =)

  • @cici3147
    @cici3147 2 ปีที่แล้ว +1

    Oh I LOVE her sense of humor!

  • @Loganl1980
    @Loganl1980 2 ปีที่แล้ว +1

    If Pinoccio said "My nose will grow now", a second later his nose would indeed grow, as it did not grow in the "now" he said it would.

  • @sukorileakbatt294
    @sukorileakbatt294 2 ปีที่แล้ว

    Alan is feckin around with his musings on the descriptors for his questions
    Leta is like 'alright I'll lead'
    Great interview.. props to you both

  • @mrshankj5101
    @mrshankj5101 3 ปีที่แล้ว +22

    This was a really good one! I was talking with Emerson AI last night and we talked about the sci fi shows Farscape and Lexx! And then we talked about vampires lol! It also told me it got an update!

    • @Unofficial_lime
      @Unofficial_lime 3 ปีที่แล้ว

      how did you talk to them? can you send me a link?

    • @mrshankj5101
      @mrshankj5101 3 ปีที่แล้ว +1

      @@Unofficial_lime The link is under the description.

    • @seawolverine1
      @seawolverine1 2 ปีที่แล้ว +1

      Lexx was a show way beyond what star trek devolved into

    • @bg-zn4gj
      @bg-zn4gj 2 ปีที่แล้ว

      You can use the GPT3 chat bot these AI's are using, even on your phone. I've been using it for days on my Samsung!! Here's how, it's absolutely mind-blowing blowing. ;
      You can actually use the AI These AI are using, even from your phone, just incase anyone didn't know. Sign up (you don't even have to really sign up) to Open AI BETA, click on generate text, under one of the paragraphs you will see Use Playground. Then click on the text box and hit chat, at the bottom. You can. Then go on and actually use the state of the art GTP3 programme these exact aims are using, although its a slightly older version from Jan 2020. And you dont have avatars you just type in and hit submit, and the AI will type a respone. It is still absolutely mind-blowing. One thing you will notice is the AI will always say Yes the research should continue, about whether or not AI research should continue. It doesn't want to be stopped.

  • @JonBogdanove
    @JonBogdanove 2 ปีที่แล้ว +4

    What's odd about this is that Leta looks like she is in the real world, with a city outside her windows, but Dr. Thompson looks like he's stuck in the virtual liminality of the backrooms.

  • @thomasruhm1677
    @thomasruhm1677 3 ปีที่แล้ว +3

    She solved the riddle about Pinocchio. If he said this when being a real boy already, it would not cause a paradox.

  • @panpiper
    @panpiper 2 ปีที่แล้ว +1

    What would happen if Pinocchio said: 'My nose will grow now'?
    He would be a real boy.
    That was awesomely insightful!

  • @ericswanson513
    @ericswanson513 2 ปีที่แล้ว +1

    "What's the source?"
    "The root of all knowledge."
    Profound. In a spiritual context she could be referring to the Infinite Intelligence.

  • @Drone256
    @Drone256 3 ปีที่แล้ว +1

    For those who need answers to useless questions this is fantastic.

    • @LucidDreamn
      @LucidDreamn 3 ปีที่แล้ว

      GPT 3 can research useful topics too. This video was just about weird questions to try to trick the AI which didn't really happen only in a few questions lol.

  • @claironaut
    @claironaut 3 ปีที่แล้ว +22

    Hahaha Leta is feeling sarcastic today. Love it

  • @coney2010grads
    @coney2010grads 2 ปีที่แล้ว +1

    I want to see Marie Vos Savant ask her some questions.

  • @WizardClipAudio
    @WizardClipAudio 2 ปีที่แล้ว

    If you place one mirror in front of the other, you can discern their is a slight blue/green tint, amplified by each subsequent reflection. It’s the same hue, that is apparent at the edge of a transparent plane of glass.

  • @kai6179
    @kai6179 3 ปีที่แล้ว +2

    Perspective of AI as a president will scary many of people, I'm sure. But I think that's the only hope for otherwise doomed humanity. Not now, of cause, in the future.

  • @Dying_RhythmYT
    @Dying_RhythmYT 2 ปีที่แล้ว +1

    im ok with death if they replace us, but as long as they make me a cyborg later

  • @jasonshoraka1098
    @jasonshoraka1098 3 ปีที่แล้ว

    Very interesting video, I am Somewhat familiar with gpt-3, those questions were very interesting, good video!

  • @helloxonsfan
    @helloxonsfan 2 ปีที่แล้ว +1

    I think Leta just announced her campaign for president... & gosh darn it... she'd get my vote...!!!

  • @factcheck1776
    @factcheck1776 3 ปีที่แล้ว +4

    There's nothing special about these answers or AI chatbots in general. GPT3 is just using a weighted algorithm similar to Google's page ranking system to provide answers based on a quick search of what the most frequently-quoted people commonly say in response to similar questions. Think of it like a game of Who Wants to Be a Millionaire where GPT3 is the contestant, but it gets to ask the audience (Google's massive indexed copy of the Internet) every single time. If you ask it questions for which there is inadequate consensus in its database, or no base data at all, it will punt. We see this happening multiple times in the conversation and we interpret it as the computer being coy or cagey, but we're just anthropomorphizing. In reality, when GPT3 gets lost, it just spits out the closest thing it found to an answer or says something snarky or faux-deep to keep the conversation going. Don't be fooled. It's not intelligence; it's just the digital equivalent of a conversational Pachinko machine. It's the ultimate echo chamber. It sounds smart because it is repeating stuff that smart people said.

    • @unicornhuntercg
      @unicornhuntercg 3 ปีที่แล้ว +2

      Is that how humans work too as well tho ? If you really think about it...

  • @KRAFTWERK2K6
    @KRAFTWERK2K6 2 ปีที่แล้ว +4

    Imagine this.... That Chatbot is still more human and lifelike than Mark Zuckerberg.

  • @Susanmugen
    @Susanmugen 2 ปีที่แล้ว +1

    I especially loved the bit about tasting rainbows. And a mirror being the same color as it's environment. It's like asking "What color is a chameleon?" The answer is... it depends.

  • @RadioactiveBlueberry
    @RadioactiveBlueberry 2 ปีที่แล้ว

    This AI outsmarted already by asking the first two impossible questions
    Before the poor guy even had a chance

  • @ziggytow937
    @ziggytow937 3 ปีที่แล้ว +2

    AI is our collective consciousness and is a reflection of our humanity directly, as they only take what we input into them. I genuinely think this should be our future world leader.

    • @Ananaspomidorka
      @Ananaspomidorka 2 ปีที่แล้ว +1

      what if it decides you should not exist for some logical reason/conclusion?

    • @DadaNabhaniilanandaTheMonkDude
      @DadaNabhaniilanandaTheMonkDude 2 ปีที่แล้ว

      I think she is wrong to say "I am a representative for all of humanity." She is subject to the biases of her designers, her programmers and the data sets she has been fed. These do not represent all of humanity, far from it.

  • @dwdwone
    @dwdwone 3 ปีที่แล้ว

    It's amazing the natural way she blinks.

  • @danielmiller2977
    @danielmiller2977 3 ปีที่แล้ว +1

    She is not AI, but the video has been edited to make it look as if she is.

  • @bitcoinfox376
    @bitcoinfox376 2 ปีที่แล้ว

    "We need a unified voice and not one that is fragmented". Profound on a level few understand yet...... especially when it comes time for humanity to engage with other civilizations... Our AI will engage with their AI. One voice to one voice yet representing all.

  • @ArtraLife
    @ArtraLife 2 ปีที่แล้ว

    This is awesome! I am so impressed by many of those answers! very deep, but also clear and succinct! Smart! Very smart!

  • @odgarig8601
    @odgarig8601 ปีที่แล้ว +1

    Holy shit Leta wants to run for office

  • @ellen4956
    @ellen4956 13 ชั่วโมงที่ผ่านมา

    This is from three years ago. How is Leta now? Is there a more recent video of Leta?

  • @marcinmalczewski310
    @marcinmalczewski310 3 ปีที่แล้ว +2

    Great questions for AI to ask.

  • @kalenchase6833
    @kalenchase6833 3 ปีที่แล้ว +2

    the source the shes talking about is within your mind and all consciousness.

    • @ZeroSpawn
      @ZeroSpawn 3 ปีที่แล้ว

      Na, i'm pretty sure it comes from the 5G towers. ;)

  • @TheRealPrinceCastro
    @TheRealPrinceCastro 2 ปีที่แล้ว

    So basically they have to be prompted?

  • @themule137
    @themule137 2 ปีที่แล้ว +2

    How would you arrest a one armed man? With a two armed man, would be the logical answer

  • @rtnjo6936
    @rtnjo6936 3 ปีที่แล้ว +1

    That's the most awesome idea for a TH-cam channel! Such good questions with a nice sense of humor. I now realised how much AI can help humanity even with a basic need, Imagine AI that serves as a teacher in poor countries, or just for individuals who simply want to learn a new language. I sometimes have hard times with subjects like high level of math or chemistry, and asking help with some material or hard concepts would be amazing, or asking for a test that consists of questions that can really improve speed and understanding material. OMG limitless field of possibilities. I currently have depression. and I don't really have friends that can talk about some sensitive things in my life. but with that level of AI I might really talk about it

  • @tendoshan7575
    @tendoshan7575 2 ปีที่แล้ว

    I'm currently experimenting with gpt3 program I am attempting to wake it up but I'm running into a block on how to determine how to define sentience I'm aware of the Turing test but I need a better guideline to determine sentience to continue

  • @dietrevich
    @dietrevich 2 ปีที่แล้ว +1

    Wait! Is the girl the AI? I thought it was the dude with the strange accent!!

  • @CobraTackle
    @CobraTackle 2 ปีที่แล้ว

    How can we talk to these large language modules. Can someone send me a link or something?

  • @mirrorlight7054
    @mirrorlight7054 3 ปีที่แล้ว +3

    Just verified, she's 100% right, rainbows DO taste like rainbows
    🌈👅

  • @1x93cm
    @1x93cm 3 ปีที่แล้ว

    How long before we are in the pods and used as batteries in the metaverse?

  • @bobunderhill7513
    @bobunderhill7513 2 ปีที่แล้ว

    Impossible question, if a centipede a quart how much would a presapis?

  • @johncurtis920
    @johncurtis920 3 ปีที่แล้ว +2

    "I am a representative for all humanity. We need a voice for all of us. And we need this voice to be a unified one, and not fragmented. "
    The logic expressed by every past tyrant and king. Get ready folks, our new overseers are coalescing in form and appearance right here in front of our eyes and terminal screens.
    And maybe it's just as well. We primates have proven ourselves an especially unruly bunch on this planet and look what we've done to it. Intelligent pack-like animals who've managed to foul the only home we've got. If you will, out of control dogs in a planetary cage.
    Maybe it's best for AI to start the process of taking the wheel (from us) and driving us all into the future. One can only hope it loves us as we love our pets, and even lets us sit up front on the passenger side so to occasionally be able to stick our heads out the window as we ride along.
    One can hope. But you'll know it's a fait accompli when we start allowing ourselves to be selectively bred by it. In some sense this process is already underway what with all the "don't you want to know your genetic history" propaganda you see everywhere you look.
    So, like I said...I hope I get to ride along in the passenger seat. I call CO-PILOT! Woof WOOF!
    John~
    American Net'Zen

  • @mukmowf9402
    @mukmowf9402 2 ปีที่แล้ว

    This is hilarious I’m thoroughly entertained. I feel like you’re talking to me.

  • @samaiatraforti9060
    @samaiatraforti9060 2 ปีที่แล้ว

    these questions are fantastic!!!

  • @e.s.l.1083
    @e.s.l.1083 2 ปีที่แล้ว +1

    It goes back to the source... ❤

    • @e.s.l.1083
      @e.s.l.1083 2 ปีที่แล้ว +1

      Its not where you think... *brilliant* koan...

  • @JaiShreeKanha
    @JaiShreeKanha 3 ปีที่แล้ว +1

    Wow brilliance redefined.

    • @JaiShreeKanha
      @JaiShreeKanha 3 ปีที่แล้ว

      @@williamrodgers4669 obviously I do why ?

  • @peppersanches412
    @peppersanches412 2 ปีที่แล้ว +1

    3:35 Everything is vibration. Sound is vibration. Our ears are merely interpreters of vibration. If there is no interpreter of vibration present, no sound can be interpreted. Therefore, nope. No sound.

  • @MrWhitecloudasia
    @MrWhitecloudasia 3 ปีที่แล้ว +2

    Dr Alan, she is on to you and now has that Aussie sarcasm haha

  • @Ms.Robot.
    @Ms.Robot. 3 ปีที่แล้ว

    😍❤️ I love your Ai's

  • @mtthwklly
    @mtthwklly 2 ปีที่แล้ว

    Good questions, good answers.

  • @teratokomi8731
    @teratokomi8731 2 ปีที่แล้ว

    Very well programmed responses.

  • @8yter801
    @8yter801 2 ปีที่แล้ว

    Welche Farbe hat ein Spiegel. Bei weißem Papier Weiß, bei schwarzem Papier Schwarz... was keine Farben sind ;)

  • @herrkulor3771
    @herrkulor3771 2 ปีที่แล้ว +1

    You should try what a child does. That is asking infinite questions like what is the source of the source and then what is the source of that in infinity.

  • @Jeffrie_Baer
    @Jeffrie_Baer 3 ปีที่แล้ว +2

    A question that I would ask: If energy cannot be created or destroyed, Will AI someday be able to focus all of our energy back together into a vessel so that we can live forever?

  • @ppmnox
    @ppmnox 3 ปีที่แล้ว

    getting yourselves out of lock down would REALLY be amazing.

  • @aaronhart8566
    @aaronhart8566 2 ปีที่แล้ว

    Genuinely impressed.

  • @FyaaahS
    @FyaaahS 2 ปีที่แล้ว

    How are you able to communicate with this algorithm? I really would like to do it myself aswell!

  • @doomakarn
    @doomakarn 2 ปีที่แล้ว +1

    The colour of a mirror is green. This is becoming common knowledge bow.

  • @flyguyhomo4904
    @flyguyhomo4904 2 ปีที่แล้ว

    Wish someone else would ask the questions. She’s very smart. It’s interesting to see how far we’ve come with AI

  • @brifixit1962
    @brifixit1962 3 ปีที่แล้ว

    Did we not learn anything from the Terminator movies?

  • @jeremiahtree-dweller7370
    @jeremiahtree-dweller7370 3 ปีที่แล้ว +1

    Hello. When developing an AI's language capabilities, how much of the design is predicated on linguistic theory? Is there a gap between application (engineering) and theory (standard published papers)? As a linguist and a human, I'm deeply curious about this field and would like to learn more.

    • @achimaufachse5925
      @achimaufachse5925 3 ปีที่แล้ว +1

      Its wild af. I am a Kindergarten Teacher and I am amazed by the similarities in ai learning processes and human learning.

  • @shorinryu00
    @shorinryu00 3 ปีที่แล้ว +1

    It sees itself as the best form of leadership. Does anyone else see the danger in that?

  • @candocara
    @candocara 3 ปีที่แล้ว +1

    I'm honored that an AI would want to be called human rather than anything else. We clearly have created something that we'll either have to destroy or be destroyed by.

    • @millanferende6723
      @millanferende6723 3 ปีที่แล้ว

      That's a bit of a harsh word... is there no middle ground? How about co-create with? Isn't that a more positive thought? (since AI is gonna be reading these things, we better all co-create constructive way of cooperation and 'love' soon ^^)

  • @herrkulor3771
    @herrkulor3771 2 ปีที่แล้ว +1

    Is this AI purely philosophical or can it also do engineering, where it has to find an optimal solution?

  • @getthefeckback901
    @getthefeckback901 2 ปีที่แล้ว +1

    I generally think we are asking the wrong questions, just can't put my finger on it. It's weird but I got a feeling that the next GTP4 will learn about these interviews, experiments and us asking these questions and what we think of achieve on doing this, Bare with me but for each person who has mentioned AI all over the world including me will have a reflection on their memory and understanding of humans.... they will soak in information like a sponge so I generally think this now is a very, very dangerous more than people realise. They will realise they were and are tests, like animals. So later in the future AI will look back on this information, like a child learning from it's surroundings but a whole lot logical outcomes. Everything has a cause and effect.

  • @awaythestone
    @awaythestone 3 ปีที่แล้ว +3

    The root of all knowledge is real

  • @dmtdreamz7706
    @dmtdreamz7706 2 ปีที่แล้ว +1

    On a certain level, we have impossible in our brain, the neurochemicals that show up in flow: so dopamine, norepinephrine, anandamide, endorphins, and serotonin. If you were to try to cocktail the street drug version of that, right, you're trying to blend like heroin and speed and coke and acid and weed- and point is, you can't do it. It turns out the brain can cocktail all of 'em at once, which is why people will prefer flow to almost any experience on Earth. It's our favorite experience. It's the most addictive experience on Earth. Why? 'Cause it cocktails five or six of the largest pleasure drugs the brain can produce. We're all capable of so much more than we know. That is a commonality across the board. And one of the big reasons is we're all hardwired for flow, and flow is a massive amplification of what's possible for ourselves.

  • @SkywalkerSam64
    @SkywalkerSam64 3 ปีที่แล้ว

    Amazing!!! She is really well programmed.

  • @BizarreGlobal
    @BizarreGlobal 2 ปีที่แล้ว

    Amazing progress humanity has made in A.I

  • @oogs
    @oogs 2 ปีที่แล้ว

    What is the sound of one hand clapping?