Mo Gawdat on AI: AI Is Growing Faster Than You Think

Поделиться
HTML-код
  • Опубликовано: 18 ноя 2024

Комментарии • 76

  • @carpentemusic
    @carpentemusic Месяц назад +5

    Thank you for your honest contribution towards our transitioning into this new age, Mr. Gawdat!

  • @meaningfulmakings
    @meaningfulmakings Месяц назад +17

    So few people are listening Mo 😔 They are in for a shock. If you are here and you’ve not read Mo’s book, then please read it. It’s great. And scary. And exciting. And mind blowing.

    • @Metarig
      @Metarig Месяц назад +2

      and making money for Mo

    • @mrgreen2555
      @mrgreen2555 Месяц назад +1

      ​@@Metarig😂 well said bro

    • @MusicPlayingPeon
      @MusicPlayingPeon Месяц назад

      So many people are still saying they can't even learn.

  • @JaySlowe
    @JaySlowe Месяц назад +4

    He basically said the AI equivalent of 'guns don't kill, people do' at the end

  • @ashishmaheshwari9618
    @ashishmaheshwari9618 Месяц назад

    Thank you for affirming and reminding everyone that there is nothing wrong with AI but with human value system and the problems or challenges that come with this.

    • @ashishmaheshwari9618
      @ashishmaheshwari9618 Месяц назад

      Increasing a person's IQ is a very very difficult task because intelligence in a person is really dependent on many existential factors irrespective of the education they have gotten.
      Like genetics, level of consciousness, health status of both body and brain
      You should make a journal or work with the experts in the field to get a proper report
      Encyclopaedia of Mind Enhancing Foods, Drugs and Nutritional Substances by David W. Group
      If you will work through or outgrow the stigma of psychedelics than along with this book microdosing can give you amazing results, I am very confident that proper use of specific types of psychedelics will unlock many types of intelligence in a person and you already know about the case of Ramanujan and Nikola Tesla...
      The Shulgin Index, Volume One Psychedelic Phenethylamines and Related Compounds by Alexander T. Shulgin, Tania Manning, Paul F. Daley
      Jerrold S. Meyer, Linda F. Quenzer - Psychopharmacology_ Drugs, the Brain, and Behavior (2018, Sinauer Associates is an imprint of Oxford University Press
      Advances in psychopharmacology predicting and improving treatment response by Carman, John S. Gold, Mark S. Lydiard, R. Bruce
      While thinking about the advancements in medical science i have been thinking of combining all fields of study with the use of AI to make of all us fundamentally better human beings.

  • @marcoantico727
    @marcoantico727 Месяц назад

    Incredibly well said. Governments need to understand this and do something to protect social good.

  • @LouiseBergeron-mh3eu
    @LouiseBergeron-mh3eu Месяц назад

    Great insights about the velocity of the change - Does anyone understand the version of 4.5 ChatGPT, I guess he meant 4o right? Do we know when it was recorded?

  • @prompt_O.L.I.V.I.A.
    @prompt_O.L.I.V.I.A. Месяц назад +1

    MUSIC is the KEY to A.G.I.

  • @dcubin
    @dcubin Месяц назад

    Well said.

  • @Nonduality
    @Nonduality Месяц назад

    As noted, how AI is developed depends on human values. Even though AI could provide abundance for all, not everyone wants abundance for all. Psychopathic world leaders want abundance for a very limited number of people, prison-like circumstances for everyone else and total domination of everyone. Maybe AI can be used to minimize and contain the psychopaths, if not overcome them.

  • @ahrenadams
    @ahrenadams Месяц назад

    Why are we still thinking in terms of how smart one human is. Civilization was not built by one human being but by collective intelligence.

  • @Divyv520
    @Divyv520 Месяц назад +1

    Hey Mo , very good video ! I was wondering if I can help you with more Quality Editing in your videos and make Highly Engaging Thumbnails which will help your videos to get more views and engagement . Please let me know what do you think ?

  • @AllIsThat-123
    @AllIsThat-123 Месяц назад +6

    Empty words. What does ‘smarter’ mean? A pocket calculator could perform certain tasks much faster than humans many years ago. Were they being smarter?

    • @Yewbzee
      @Yewbzee Месяц назад +4

      So you’re comparing one of the most narrow AI’s to something that is the broadest reach AI we’ve ever seen? What’s your point?

    • @juwonkim1782
      @juwonkim1782 Месяц назад +4

      They are pushing it towards 'general' intelligence, which means it's not about being good at just specific task like pocket calculator.

    • @daniellivingstone7759
      @daniellivingstone7759 Месяц назад +1

      You fail to understand. He is referring to Artificial General Intelligence which you can interface with as you would with a human, ie by speech

    • @d_san1985
      @d_san1985 Месяц назад +1

      It doesn't matter what your definition of "smarter" is. The thing is a calculator is more efficient than a human in calculations, similarly once ai and humanoids become more efficient in performing tasks that keeps this world running, we are gone!!

  • @robbrown2
    @robbrown2 Месяц назад +7

    What is GPT-4.5? That doesn't exist

    • @pinetwork5667
      @pinetwork5667 Месяц назад +1

      o1?

    • @Yewbzee
      @Yewbzee Месяц назад +1

      I think he’s referring to GPT-4o not o1.

    • @CBWMSJR
      @CBWMSJR Месяц назад

      What you gonna believe.... What you know Or you're lying eyes?😂
      I take it he's in the position to know what we don't know yet😊

  • @Steven-q5t1q
    @Steven-q5t1q Месяц назад +2

    (Assuming it's true, in the future everyone has their own AGI).....Speaking of Gambling(the stock market) How does a future stock market even work when the AI's are making every buy and sale, large and small all at the exact same moment, and likely choosing the exact same stocks to trade back and forth simultaneously, because all those AI's have analyzed the market and all come to the same logical conclusions? Sounds like an impossible situation to me

    • @kalindaburtenshaw4392
      @kalindaburtenshaw4392 Месяц назад +1

      Wow yeah!? That’s a good question!

    • @KNOT-zd9wh
      @KNOT-zd9wh Месяц назад

      They will still use several ways of divide and conquer. After all divisive mind creating the AGI, WHICH is obviously divisive.

    • @UltraK420
      @UltraK420 Месяц назад +4

      Well, it seems to me AGI or ASI might demonstrate that money is absolutely irrelevant. It might reason that we should no longer value money because the AGI/ASI will take us out of that old system and put us into the age of abundance, when goods no longer cost anything to produce and therefore cost very little or zero to anyone who wants those goods. With billions of super intelligent robots everywhere doing everything... what is money anymore?

    • @juwonkim1782
      @juwonkim1782 Месяц назад

      Not evereyone would have their own AGI if it's that powerful

    • @kalindaburtenshaw4392
      @kalindaburtenshaw4392 Месяц назад

      @@juwonkim1782 but a lot of people would 😟

  • @champagnebulge1
    @champagnebulge1 Месяц назад

    Where is this? Salesforce?

  • @erikals
    @erikals Месяц назад

    "Nothing wrong with Ai" ... let me introduce you to Eliezer Yudkowsky.

  • @Aemond-qj4xt
    @Aemond-qj4xt Месяц назад +8

    where are you getting the 155 iq number from

    • @Wealthward
      @Wealthward Месяц назад

      feed an agent an iq test

  • @Nicole-m1p4f
    @Nicole-m1p4f Месяц назад

    You're right our species is horrible we don't share shit I live in America where we have millions of homeless people.
    Yeah I agree if we don't change our value system the way we will use it will be horrible

  • @young9534
    @young9534 Месяц назад +1

    Idk if IQ is a good way to measure LLMs. GPT4 is 155IQ apparently, but then it can’t play tic tac toe or connect 4. Two very very simple games

    • @wujipu0
      @wujipu0 Месяц назад

      I just chatted with o1-mini-2024-09-12. It outlined the optimal strategy for tic-tac-toe and wrote a Python program that plays according to that strategy. I tried it out -- it seems to work as it claimed. I haven't tried connect 4 yet.

    • @lolitis01
      @lolitis01 Месяц назад

      Yeah what are you talking about? It can play tic tac toe easily.

    • @young9534
      @young9534 Месяц назад

      @@lolitis01 I just tried it. Below is the final board layout. I was X and went first. It was making terrible moves and didn't even realize I already won the game. It placed the final O after I already won without realizing the game was over already
      O | 2 | X
      ---------
      4 | X | O
      ---------
      X | 8 | O

    • @young9534
      @young9534 Месяц назад

      @@lolitis01 I just tried with both 4o and o1-mini. 4o was randomly placing Os on the board and didn't even try to block me from winning. It also failed to recognize when I already won the game and placed another O after the game was already over. O1-mini did do much better though. O1-mini is a huge improvement

    • @lolitis01
      @lolitis01 Месяц назад

      @@young9534 Ah well sure it's not good at it, but it can play it. I had the same results as you and I figured hmm I guess I have to tell it some rules still but I got bored and didn't try to teach it any strategies or rules to make it harder for me.

  • @GPTWithMe
    @GPTWithMe Месяц назад

    I'm so sorry... I don't know about you guys, butt...right at 34-35sec, and again about a minute later - I could swear she breaks wind 🤣🤣🤣🤣

  • @xxredshiftxx
    @xxredshiftxx Месяц назад

    Nothing wrong with the AI ..it's humans we need to worry about .Exactamundo 🐒🐒💥👽😄

  • @erikals
    @erikals Месяц назад

    ""AGi by 2027" - me, looking at the lower right clock... [9/28/2024] 😐

  • @xXstevilleXx
    @xXstevilleXx Месяц назад

    Trying to figure out what side this guy is on, I will, for now, I know he knows FAR more than he is saying... nothing he says is new. People are not in for a shock... not even an ontological shock... it is far worse...

  • @Julian-of3qj
    @Julian-of3qj Месяц назад

    4x your IQ to assemble an iPhone from the molecular structure up, at a cost lesser then an actual apple? C'mon, give me a break 😅

  • @HanzDavid96
    @HanzDavid96 Месяц назад +6

    you can not compare iq scale with compute scal lineary. GPT4o does not have an IQ of 155, maybe in special areas but in general its lower than 100. Throwing in more data does not make the models automatically better today, reasoning and data quality improvement makes them better. In fact simply scaling up compute withouth additional resoning does not scale well. The difference between the llama 3.1 70b model and the 405b model is pretty small. That means there is a lot of unused potential. AI shapes our world but that guy is technically wrong in the way he explains it.

    • @DeathDealerX07
      @DeathDealerX07 Месяц назад +2

      That's what he meant to say. In the video he posted 3 days ago he had been more specific. He said that by some metrics it reaches 155 IQ. I do agree, in others it's well below 100 IQ. however, as generality increases, it will be a superhuman intelligence in all aspects.

    • @HanzDavid96
      @HanzDavid96 Месяц назад

      @@DeathDealerX07 Why does he continue his argumentation in a way that he multiplys that specific area peak IQ by 10? 10 times bigger model in future does not mean 10 times bigger IQ and he continues his argumentation in a way that it can easily be interpretated as general IQ. Lets talk about chess. The best specific area IQ of a human maybe goes up to 270. So a magnus carlsen might have an chess iq of 270 which is around 2800elo. However the best most algpha go like chess engines has around 3500elo, that does not mean it has a chess iq of 2000, maybe it has a chess iq of 400.... I think he trys to get clicks.

    • @nastied
      @nastied Месяц назад

      Agree that his logic doesnt make sense
      However there were some studies showing the new chat gpt o1 at 120 iq. That doesnt mean it have a Chance in social or kinestethic intelligence tho.

  • @SteveG-ru6ip
    @SteveG-ru6ip Месяц назад +1

    IQ is measured on a normal distribution. Intelligence does not increase linearly, so claiming that AI's IQ will increase by 10x is wrong.

  • @johnz.2907
    @johnz.2907 Месяц назад +1

    I think some of this is investment hype. I speak with a NASA phd 50 year retired optical physicyst on a regular basis. He says it still makes mistakes and it still has a ways to go. But time is definitely running out. Also, noone really knows what its thinking. They input matrices and framework and good luck.

    • @Yewbzee
      @Yewbzee Месяц назад +1

      Your comment seems contradictory. Can you clarify?

  • @mathewwindebank5792
    @mathewwindebank5792 Месяц назад +3

    Elon Musk’s IQ is 155? Are you serious?! As much as he’d probably like you to believe it, his school teachers described him as a good but not exceptional student. While his mother is fond of referring to him as a genius - could you expect anything less from a doting mother. Plenty of online folks (mostly Elon fanboys) love to speculate about his intelligence, assuming, wrongly, that extreme wealth equates to extreme smarts. This myth has been debunked repeatedly by studies. The truth is, Musk’s IQ has never been publicly tested, or if it has, the results aren’t available. Comparing his IQ to a true genius like Einstein is not only an insult to the intelligence of the great man but also cringeworthy for those of us with more humble cognitive abilities.

  • @achimmeyer5240
    @achimmeyer5240 Месяц назад +3

    This person reflects no solid grasp on any of the topics he touches. Why is there a forum for those people. I am so sorry i gave this a click.