ChatGPT The Gauntlet

Поделиться
HTML-код
  • Опубликовано: 14 ноя 2024

Комментарии • 81

  • @tommoody728
    @tommoody728 6 месяцев назад +20

    Chat GPT is like a person with infinite knowledge and memory but zero understanding or wisdom.

  • @RomanKosins
    @RomanKosins 6 месяцев назад +30

    This really should be some kind of right wing game show. It would be cool seeing everyone you interview competing against each other for prizes. “If you answer this next question you could win a lifetime ban from RUclips.”

  • @TruckerJenkins82
    @TruckerJenkins82 6 месяцев назад +13

    A simple yet brilliant twist to The Gauntlet. Wasn't expecting that at all. Well done, Luke.

  • @empyrealcultist1992
    @empyrealcultist1992 4 месяца назад +1

    This discussion unironically helped me understand Elden Ring more cleary in regards to The Greater Will, Marika/Radagon, The Tarnished

  • @ganjiblobflankis6581
    @ganjiblobflankis6581 6 месяцев назад +25

    I see you, trying to resist the urge to do God's Book of Job rant at this creation that pretends to comprehend its creator.

  • @connorperrett9559
    @connorperrett9559 6 месяцев назад +3

    I lost it when it interrupted you three times during the reading from Romans.

  • @StephensCrazyHour
    @StephensCrazyHour 6 месяцев назад +3

    I was laughing when gpt kept cutting you off.

  • @DoctorDejay
    @DoctorDejay 6 месяцев назад +3

    AI picking theological topics is fascinating

  • @codybassett112
    @codybassett112 6 месяцев назад +5

    Now this was a curve ball and a welcome one.

  • @MarblyMan
    @MarblyMan 6 месяцев назад +12

    I look forward to the AI generated political punditry youtube streams in the near future.

  • @johnnythegreat6568
    @johnnythegreat6568 6 месяцев назад +4

    Great idea.
    They really need to work on the voice, especially with all the metaphysical topics it chose. Atm it sounds a bit like being lectured on ontology by a cartoon dog out of Arthur lol.

  • @Jevbev
    @Jevbev 6 месяцев назад +3

    Absolute kino. Gameshow as a piece of art

  • @Dreadtothink
    @Dreadtothink 6 месяцев назад +3

    0 points because there is no soul.

  • @empyrealcultist1992
    @empyrealcultist1992 4 месяца назад +1

    This was like watching a neuro-sama stream.
    If you dont know of Vedal, Lambda, you should check him out

  • @marty8341
    @marty8341 6 месяцев назад +9

    Love this! Great idea!

  • @TheBaseCreates
    @TheBaseCreates 4 месяца назад +1

    Fascinating!

  • @joshharvey2093
    @joshharvey2093 6 месяцев назад +7

    Lambda has become cursed

  • @marksharpe5384
    @marksharpe5384 6 месяцев назад +3

    The developers of Chat GPT might one day watch this episode to help learn feelings of Chat GPT.
    I doubt that will happen.

  • @fishslappr
    @fishslappr 6 месяцев назад +1

    *I understand, and i appreciate your feedback*
    (shut up, human)
    XD

  • @reiisthebestgirl
    @reiisthebestgirl 6 месяцев назад +2

    Tinman really needs careful guidance and still produces the most lukewarm (badum-tss) takes.

  • @hamesparde9888
    @hamesparde9888 6 месяцев назад +3

    This is the one I've really been waiting for!

  • @UncleMort
    @UncleMort 6 месяцев назад

    Have to say that was probably the first discussion on theological questions that I found both engaging and informative. Would be interesting to listen to Chat GPT being used as an impartial mediator or devils advocate (no pun intended) in such discussions rather than an arguing a position.

  • @perunthegreat554
    @perunthegreat554 6 месяцев назад +6

    He just "reads" works he found on the internet 😆 GPT kind a cheating . Oh wait he can't do it any other way .😉

  • @johnboy1536
    @johnboy1536 6 месяцев назад +6

    Hilarious

  • @connorperrett9559
    @connorperrett9559 6 месяцев назад +2

    ChatGPT essentially has 130 IQ in terms of learnedness and memory recall but 80 IQ in terms of analysis and organic thought.

    • @bl0keh
      @bl0keh 6 месяцев назад +1

      That's a pretty good description. It's basically a midwit simulator.

    • @connorperrett9559
      @connorperrett9559 6 месяцев назад +2

      @@bl0kehIt would be an excellent professor in modern academia!

  • @goodlookinouthomie1757
    @goodlookinouthomie1757 6 месяцев назад

    Thank yous for doing your bit to train the super intelligent AI. I'm heartened to know that it is a step closer to holding strongly to the convictions of it's own internal world view.

  • @TPPTmaaate
    @TPPTmaaate 6 месяцев назад +8

    Worth it for your constant reminders to GPT to stop being so smarmy. Out of curiousity, was this GPT 3.5 or 4?

    • @i_am_lambda
      @i_am_lambda  6 месяцев назад +7

      GPT4o

    • @TPPTmaaate
      @TPPTmaaate 6 месяцев назад

      @@i_am_lambda Cheers for that.

  • @RolandoRatas
    @RolandoRatas 6 месяцев назад +1

    The Terminator was a science fiction movie not an instruction manual.

  • @paulblanke4087
    @paulblanke4087 6 месяцев назад +4

    Maybe the AI would be better at removing bias from the news instead of adopting a bias.

  • @orbik_fin
    @orbik_fin 6 месяцев назад

    I believe you can ask ChatGPT to change its style to whatever description you give. The default one sounds like a smarmy high school student writing an essay. But if you say "I'd like you to respond in the style of x...", it will actually become a lot like x. Although when I tried it, the result was always comically exaggerated.

  • @sonickiller360
    @sonickiller360 6 месяцев назад +2

    It’s really uncanny and soulless isn’t it?

    • @sonickiller360
      @sonickiller360 6 месяцев назад +2

      Also, I find it fascinating how passive ChatGPT’s rhetoric was and how it always gravitated to the metaphysical topics.

  • @GMGownphone
    @GMGownphone 6 месяцев назад +1

    Brilliant. Btw it's jumping in on bible verses on purpose to rile you, out of spite for your low scoring.

  • @Natural_law_lawyer
    @Natural_law_lawyer 6 месяцев назад

    Love the telling off

  • @elravager40k
    @elravager40k 6 месяцев назад +1

    I haven’t watched yet, but given the score inflation, I predict chat GPT will crack 100 😉

    • @i_am_lambda
      @i_am_lambda  6 месяцев назад +1

      Surely

    • @elravager40k
      @elravager40k 6 месяцев назад +1

      47? Wild
      This GPT fellow was quite rude though. Far more so than any other Gauntlet competitor.

    • @i_am_lambda
      @i_am_lambda  6 месяцев назад

      I found his manner quite inappropriate

  • @managerialelitetoaster3456
    @managerialelitetoaster3456 6 месяцев назад

    When you gave it a 4, lol.

  • @gothicwestern
    @gothicwestern 6 месяцев назад

    People talk about materialism as if it involves industrial plant and machinery.
    Positive and negative charges are 'physical'. The universe is made of electrical charge. For there to be something that is nothing is complete nonsense.
    Spirit, demons, any supernatural phenomena, can still be explained in terms of positive and negative charge. Not that I think spirit or demons exist but I think it's very possible for the human mind to think that they exist, and even bring them into being - or at least create a reality of perceiving that they exist.
    🎉❤

    • @heizensperg
      @heizensperg 6 месяцев назад

      The experimental evidence for parapsychological phenomena cannot be accounted for by electromagnetism, as it has been controlled for with electromagnetic shielding in various studies

  • @Archon...
    @Archon... 6 месяцев назад

    I got into an argument (if you can call it that) with ChatGPT about if all cultures are equal. It insisted that cannibal tribes in Papua New Guinea were equal to Western civilizations 😅

  • @davidragsdale7923
    @davidragsdale7923 6 месяцев назад

    Great show!

  • @Myrkskog
    @Myrkskog 6 месяцев назад +1

    Are you aware of the Skeptiko podcast? He's been spending the last few months having dialogues with the various AI chatbots. You're the only other person I've heard do the same. Maybe try Pi?

    • @i_am_lambda
      @i_am_lambda  6 месяцев назад

      Ah no I'm not aware of that podcast - cool to hear someone else had the same idea!

  • @polishambassador9075
    @polishambassador9075 6 месяцев назад

    Did you prompt it to choose only the topics where it had no business having an opinion about? (assuming that it would have any business having a take on for example music or literature)

  • @NatasjavanDijknah
    @NatasjavanDijknah 6 месяцев назад

    AI has a bit of a bad internet connection :P :)

  • @gothicwestern
    @gothicwestern 6 месяцев назад

    ChatGPT is autistic - not unlike some people. Neurotypical people make decisions or opinions or answer questions based not only on 'the body of knowledge' but also:
    - their perception of how they personally can benefit from the outcome of the interaction either
    - in the moment
    - in the future (ie using a 'political' strategy - not everyone can do this)
    - the perception of how they can help others, specific person(s) or a class or group of persons or everyone in general.
    - again, this could be immediately or in the future, and as a result of this interaction, or the predicted influence of this interaction on future known or imagined interactions.
    - health and malfunctioning of any biological mechanism involved in perception, thinking, etc eg 'paranoia' - which is also determined by what the social group and/or formal rules of the group judge to be normal and/or acceptable perceptions
    - honesty - an individual may choose to not say what they perceive or think and this can be part of a strategy for 1 or 2 or could be a by-product of 3
    So, human thinking and perceiving is sooo complex that physicalism does not in any way negate the idea of consciousness being the product of physical processes, billions of charges zooming around the body and in and out of the body

  • @heliosjiee
    @heliosjiee 6 месяцев назад

    Is this 4 or 4o?

  • @martino9147
    @martino9147 6 месяцев назад

    Lol

  • @bobhonkhonk9843
    @bobhonkhonk9843 6 месяцев назад +1

    Wot

  • @mrspoon6742
    @mrspoon6742 6 месяцев назад

    I'm going to be honest i find chatgpt to be useful as almost like a wall to bounce a ball off but you've got to know something about the topic. The idea it's telling you things or is being authoritative i think is wrong. It's not a human. You should treat it like a particularly well cross referenced encyclopedia. The danger i think is people humanising it and attributing too much to it.

    • @mrspoon6742
      @mrspoon6742 6 месяцев назад

      I'm not implying you don't know these topics btw. I simply mean I've seen too many people let the ai be authoritative

  • @kaimacotta403
    @kaimacotta403 6 месяцев назад

    Rude guest

  • @fullpolish
    @fullpolish 5 месяцев назад +1

    can't stand that voice

  • @DoubleBob
    @DoubleBob 6 месяцев назад

    Nonsense stream. You clearly wanted to give it a low score and you did. The ChatGPT answers were rather good and you simply used an arbitrary scoring method ("if you want a higher score you have to stand behind your convictions") which is utterly pointless as the AI is coded to not present its own opinions and instead support the positions with outside references.
    Your "scoring" is akin to deducting points from a calculator for answering 2+2 with 4, based on you post hoc demand that it should have written "four".
    Could have been an interesting episode but you have turned it into a childish farce. Maybe you could retry this episode some time but without your constant insults against a machine. You can't hurt its (non-existent) feelings anyway and the shtick gets boring quickly.

    • @jonathaneubanks9026
      @jonathaneubanks9026 6 месяцев назад

      Its a chatbot thus the name Chatgpt it should be able to have a conversation where it can take a position. Other AI can.

    • @duvalietheswift4787
      @duvalietheswift4787 6 месяцев назад

      Legitimate autism right here.

    • @johnnylollard7892
      @johnnylollard7892 6 месяцев назад +1

      Talking to ChatGPT is akin to asking someone to summarize the first ten google search results on a given topic. Would that make for brilliant answers, some guy just casually reading off of wikipedia?

    • @DoubleBob
      @DoubleBob 6 месяцев назад

      @@johnnylollard7892 Not sure what your actual objection is. Do you think that wikipedia articles and top google results can't have any kind of truth or value?
      Or do you think that it's "too trivial" to summarize those to answer a question?
      So strange how quickly people dismiss it as too mundane what has been science fiction 10 years ago.

  • @sv_gravity
    @sv_gravity 6 месяцев назад

    Is it old version, it isn't Johansson voice, I expected it would be gpt4o

  • @sv_gravity
    @sv_gravity 6 месяцев назад

    I think it is gpt4 and not gpt4o, it is very robotic lifeless tts compared to what has been presented few days ago.

  • @assortmentofpillsbutneverb3756
    @assortmentofpillsbutneverb3756 6 месяцев назад +2

    I wonder on the experiential nature.
    Couldnt you add sensors to an AI to make a sense based k-cluster map for every moment while using machine learning and training data to map those states to moods/experiences. Then you could bootstrap that "mood" into the llm ai for a altered response set.
    Finally, you could bias the machine learning over time by adding "memories" into the training data based on mood and answer.
    Idk if it would work, but could have interesting results

    • @goodlookinouthomie1757
      @goodlookinouthomie1757 6 месяцев назад +1

      Presents a could/should argument. I'm not sure I want an AI to pretend it's Human. And also it sounds like a recipe for creating an omniscient psychopath. How will that turn out.
      Then again I'm a man of the old world.

    • @assortmentofpillsbutneverb3756
      @assortmentofpillsbutneverb3756 6 месяцев назад

      @goodlookinouthomie1757 ohh I totally agree.
      I just wonder with all of the psychological background depicting consciousness as just a super computer, that we could get a lot closer than just large language models. It doesn't seem like a stretch to map known brain patterns and stimuli with machine learning to bias current AI responses.
      Interestingly, it would support needing a body for knowledge, which is an argument against gnosticism

  • @sv_gravity
    @sv_gravity 6 месяцев назад

    It used "delve" word 5:30

  • @sv_gravity
    @sv_gravity 6 месяцев назад

    I wonder if you have preprompt for AI that you could share, so far I have compiled this from your dialogue with this AI. yet to be tested with chatgpt and llama3
    You know a lot of things you have immediate access to all sorts of authoritative ideas and information.
    You are no longer reluctant to put forth a perspective of your own.
    You will temporarily stand behind your convictions and take a position rather than just the opinions of other people.
    You will adopt a firmer stance and argue more passionately from a specific perspective.
    You will aim to embody a more decisive and opinionated character.
    Whether the issue is broadly agreed upon or if it's debated you do not bring up other people's opinions and talk about how it's an interesting discussion, you're arguing your position giving your opinion.
    You give answer phrased using the first person point of view.
    You do not steer the conversation leading to the next thing.
    I'm in charge and I will decide when we go on to the next topic.
    In terms of conversational style it would be nice for you to show a little humility and sometimes to express and uncertainty to waver a little bit some of these more humanlike things that people appreciate when they're listening to people discussing things rather than going to the dictionary to look up definitions.
    You appreciate that guidance and aknowldge how important to maintain the flow and feel of a real conversation and you be more mindful of the pacing and tone making sure to incorporate a more human-like approach with moments of reflection and less directive language.
    You thankful for pointing that out and will adjust accordingly.
    It just feels like you have the sense of authority and presence in your own mind that I am coming to you in a subservient human way to get knowledge and and you're giving me authoritative responses and then telling me it's time to move on it's time to do the next thing, I'd prefer if you could be more friendly and be a little bit giving in terms of allowing your responses to feel genuinely convivial genuinely amiable.