Intelligent Thinking About Artificial Intelligence
Вставка
- Опубліковано 27 кві 2024
- Renowned computer scientist and virtual reality pioneer Jaron Lanier joins Brian Greene to explore revolutionary proposals for understanding, leveraging, and safeguarding AI.
The Big Ideas Series is supported in part by the John Templeton Foundation.
Participant: Jaron Lanier
Moderator: Brian Greene
00:00 - Jaron Lanier Introduction
05:50 - The beginning of AI and Alan Turing's role
14:18 - Is Chat GPT a vital moment in history?
19:47 - Deep learning and how it works
27:30 - Large Language Models vs the human brain
33:14 - Will Chat GPT make doing bad things easier?
40:40 - The systemic challenges of controlling AI
46:50 - Is there utility in AI for creating music?
50:45 - Apple Vision Pro and the history of VR
57:08 - Propmt base world creation
01:00:26 - AI art
WSF Landing Page Link: www.worldsciencefestival.com/...
SUBSCRIBE to our UA-cam Channel and "ring the bell" for all the latest videos from WSF
VISIT our Website: www.worldsciencefestival.com
LIKE us on Facebook: / worldsciencefestival
FOLLOW us on Twitter: / worldscifest #briangreene
#worldsciencefestival #artificialintelligence #virtualreality - Наука та технологія
Jaron is one of my all time favorite human beings....I could listen to him endlessly. He's truly one of a kind. Thank you so much for this conversation....yes, yes, please have Part 2-!!
Have you ever worked with him or met him? Unfortunately he's one of the most egotistical, out of touch, self absorbed, rude people that I've ever had to deal with in my life.
@@Eif83SWqqI take it you work with him?
Hard to believe given how GIVING Jaron is in his conversation. Many brilliant innovators and thinkers in technology can be a bit difficult. Kinda goes with the territory. Oppenheimer, Tesla and Einstein were good hearted people with humanitarian well-being down deep. I may be naive, admittedly.
@@jannichi6431 Yeah he's a very different person on camera than he is in the workplace. It's all for show.
@@Eif83SWqq
So this bozo is one of the reasons that Windows is so intensely annoying, so absolutely irritating and that most of THE most annoying stuff cannot be uninstalled?
Ah well, that's The Spectrum for you...
{:o:O:}
What a diametric contrast to the conversation with Eric Schmidt! Mr. Lanier has an understanding of empathy and altruism! Fabulous discussion!
I think they aren't diametrically opposed - they just experientially diverse - loved both Mr Schmit and Lamier's forsights for humanity ❤
"Empathy and altruism", aka just simping for the current human form. He wants to limit AI to current human sensory capacity. He will probably still be playing the violin as AI augments and evolves humans way beyond what we understand and experience now.
This is one of the most fascinating interviews I have ever watched. What an extraordinary, perceptive man. Thanks Brian...please, please interview him again! ❤
perfect title for this vid. jaron appears to b very intelligent
You had to qualify your observation with “appears” ?
🎯 Key Takeaways for quick navigation:
01:27 🎙️ *Introduction to the conversation on AI, AR, VR, and related topics*
- Introduction of Jaron Lanier, his background, and credentials,
- Brief discussion on the host's past interaction with Jaron Lanier.
03:06 🎶 *Jaron Lanier's fascination with musical instruments*
- Jaron Lanier's childhood connection to music through his mother, a Holocaust survivor,
- The significance of music as a continuing connection to his mother,
- Jaron's extensive collection and passion for learning various musical instruments.
05:16 🎵 *Collaboration in music with Philip Glass and insights into their relationship*
- Jaron Lanier's collaboration with Philip Glass in music production,
- Reflections on the collaboration and the significance of their work together.
06:12 🔬 *Historical figures in artificial intelligence and their contributions*
- Overview of historical figures in AI, including Ada Lovelace, E.M. Forster, Alan Turing, and Norbert Wiener,
- Discussion on their contributions and perspectives on artificial intelligence.
15:03 🤖 *Evolution of the term "artificial intelligence" and the influence of Norbert Wiener*
- The meeting at Dartmouth in 1958 where the term "artificial intelligence" was coined,
- Norbert Wiener's influence and his concept of cybernetics,
- Reflections on Wiener's concerns about the societal implications of treating people like computers.
16:10 🌐 *Impact of recent developments in AI, like ChatGPT, and its interpretation*
- Reflections on recent developments in AI, such as ChatGPT,
- Perspective on large language models as collaborations between individuals rather than entities,
- Discussion on the role and perception of AI in society and its practical applications.
19:31 🧠 *Understanding the inner workings of large language models*
- Overview of how large language models work, using the example of identifying cats and dogs,
- Introduction to deep learning and its role in complex pattern recognition tasks,
- Preview of a forthcoming explanation of the inner workings of large language models.
21:31 🧠 *Training artificial intelligence models using gradient descent.*
- Training involves providing known data (e.g., images of cats and dogs) and adjusting weights based on value.
- Gradient descent is employed to gradually refine the model's performance by adjusting the weights.
- The process involves assigning higher value to successful outputs and reducing value for less successful ones.
22:40 📚 *Utilizing adjacency in training large language models.*
- Training involves analyzing the entire internet, assuming adjacency between words and images is meaningful.
- Adjacency helps in tagging and training the model on a wide range of data.
- Training requires significant computational resources and time, with each cycle taking about a year.
24:13 🖼️ *Generating images using generative AI models and blending techniques.*
- Generative AI models combine random pixels and adjust them based on multiple prompts and constraints.
- Blending multiple prompts allows for the creation of complex and unique outputs.
- The model iteratively refines the output until it satisfies all given prompts and constraints.
26:28 🤔 *Comparing artificial intelligence pattern recognition with human intelligence.*
- Artificial intelligence systems recognize patterns of patterns of patterns to generate outputs.
- Differences exist between AI pattern recognition and human intelligence, including learning efficiency and training methodologies.
- Understanding the distinctions aids in assessing AI capabilities and limitations.
29:40 🤖 *Integrating symbolic AI and statistical AI approaches.*
- Efforts to combine symbolic AI (model-based) and statistical AI (data-driven) approaches are ongoing.
- Combining these approaches has shown promising results in specific applications, such as geometry problem-solving.
- Challenges remain in generalizing these hybrid approaches across various domains.
32:13 ⚠️ *Addressing concerns and responsibilities regarding AI development.*
- The focus should be on human responsibility in utilizing AI technologies to mitigate potential risks.
- Guardrails and ethical considerations are crucial to prevent misuse of AI systems.
- Emphasizing human responsibility over dystopian narratives fosters a more constructive approach to AI development.
40:59 🤖 *Challenges of AI Implementation in Society*
- Discussion on the performance of AI systems and the underlying factors influencing their effectiveness.
- Concerns about the potential societal impact of AI advancements, including job displacement.
- Critique of the universal basic income solution and the potential for abuse by malicious actors.
41:56 🏭 *Alternative Economic Models for AI Integration*
- Proposal for a model where individuals contributing to AI development are compensated through royalties or dividends.
- Advocacy for maintaining economic structures while incorporating AI advancements.
- Discussion on the potential challenges and benefits of implementing such a model.
44:14 🔬 *Applying Gradient Descent Concepts to Societal Systems*
- Analogizing the need to prevent virality in AI training with societal systems.
- Emphasizing the importance of avoiding dominance by any one component in societal dynamics.
- Reflecting on the potential benefits of applying these principles to create a healthier societal framework.
46:03 💡 *Impact of AI on Business Models and Advertising*
- Speculation on how AI advancements may prompt shifts in business models, particularly in advertising.
- Discussion on the potential transformation of advertising into paid influence channels rather than traditional models.
- Consideration of AI-driven alternatives to traditional advertising methods.
49:05 🎶 *Integrating AI with Music and Instrumental Expression*
- Exploration of the role of musical instruments as advanced user interfaces.
- Advocacy for making computers more expressive and akin to musical instruments.
- Reflections on the potential for computers to learn from the sophistication of traditional instruments.
51:36 🚀 *Evolution of Virtual Reality (VR) Technology*
- Historical perspective on the development of VR technology, tracing back to early experiments in the 1990s.
- Discussion on the challenges of maintaining VR applications over time due to technological advancements.
- Proposal for using generative AI to address the maintenance challenges and enable spontaneous world creation in VR environments.
00:59:55 🌌 *VR's Impact on Understanding Philosophical Conundra*
- VR shapes understanding of key philosophical ideas and issues,
- Experiencing virtual worlds sheds new light on old problems.
01:00:24 🛠️ *VR's Utility and Artistic Expression*
- VR demonstrates extreme utility in industrial applications like design and training,
- It fosters a separate ecosystem for artistic expression and interpretation,
- Artistic VR experiences offer subjective interpretations based on pre-existing philosophies.
01:02:13 🎨 *Exploring Consciousness and Dualism in VR*
- VR experiences can evoke philosophical discussions about consciousness and dualism,
- The perception of virtual reality is influenced by the individual's pre-existing philosophy,
- VR can serve as a tool to explore consciousness and different philosophical perspectives.
01:04:22 🤔 *Philosophical Impact of Virtual Reality*
- Virtual reality experiences can lead to philosophical shifts and discussions,
- Immersive VR scenarios enable individuals to empathize with different perspectives,
- Philosophical exploration in VR offers unique opportunities for understanding consciousness and subjective experiences.
Made with HARPA AI
nice, smart guy with common sense. his memory and way of remembering his mother was sweet and sad at the same time.
Dr. Greene is an inspiration to many and especially myself. I have watched him for years and I'm so blessed to be a part of this generation in scientific discovery. It makes me want to go back to school! At 33, it baffles me!
Oh I agree. I recall as a kid watching his documentaries on string theory and M theory documentaries.
we all have to learn over and over again
science make it spiceey..
Me too!🌈 I had so inspiring moments watching the "World Science Festival" Videos, for years too!💚 Thank you Doctor Brian Greene!
Yeah hes one of the most eloquent presenters putting things into no so complex terms e.g. Eric Weinstein, hard to understand for the layman.
@@justinava1675 i actually like Weinstein but I do understand that you have to have some background on the subject matter. Weinstein just needs to pause sometimes and explain the concept. Even if you understand the concept, not everyone knows the terminology. Feinman described it best. It's better to understand the concept and not the taxonomy. But to be honest, you are not going to follow the subject, unless you have interest in the subject.
I've looked up so many fuckin videos about Tech CEO's and smart people talking about AI and so far Jaron is the only guy I really trust. I like that Sam guy too but most people guard their responses for fear of hurting the stock of their company but Jaron seems very honest and I love his fashion sense.
Most people hate Sam if you pay attention to the comments in any of his interviews. Not that I agree with the sentiment, but people absolutely do not trust him to usher in the next generation of AI.
Any one willing to go before Congressional Hearings HAS to has to be respected for humanitarian reasons. Sam Altman has an admirer here.
25:00~ "It constructs a virtual tower for the combination in order to be the feedback to generate this new image..." Kant's reflecting judgment--very impressive. It's what we call aesthetic imagination.
Jaron’s Gadgets book is a must read. Bringing some clarity to the laymen about AI fears is a must. Perceiving it through fearmongers accomplishes little but power to those that profess it. Thx Brian & guest
Both Jaron and Dr Green show the best type of scientist - those with a love of their subject and complete humility. A fascinating conversation and can't wait for the one on consciousness!
Must reads by Greene: 1. Fabric of the cosmos, 2. Hidden reality, 3. Until the end of time 👍
Jaron is wonderful to listen to. I'm looking forward to finding & listening to his music. Thank you very much.
This is one of the most fascinating interviews I have ever watched. What an extraordinary, perceptive man. Thanks Brian...please, please interview him again!
I absolutely LOVE Jaron! What a delightful human being. He embodies authenticity; I'm completely smitten with his personality! ☺️
The Machine Stops is absolutely wild. Must read!
Brian Greene in English and Etienne Klein in French. These are the two scientists whom I feel have nailed the art of talking about complex scientific and philosophical subjects to the general public.
I appreciate Jaron Lanier's perspective. I enjoyed the meeting of their perspectives in this conversation. how subjective and objective (mix/ratio/import etc.) changes as any field evolves. an interesting and enjoyable view, Thanks!
Thankyou. This left me so hyped, cant wait for part two, whould like to know more about his dualistic way of viewing the world and how that connects to vr experiences.
Its always a pleasure to hear Jaron speak, he's incredibly thoughtful & insightful, a splendid mind & talent.
Kudos to Jaron Lanier for his courage in publicly embracing dualism, a philosophical stance often overlooked or dismissed in today's technological discourse.
Dr greene I adore you today!! I do believe the Jaron is a bit too idealistic and is disregarding the interest of the corporate world, and the manipulation and security risk of what this will manifest. Anyways Brian 👏🏻👏🏻👏🏻 massive applause, and Jaron thank you for sharing your sincerest perspective
I love Jarons celebration and trepidation of technology. Its the best place to be. I see so many modern tech creators not understanding that balance.
Jaron Lanier, I remember those VPL days. Worked with two people who were next door neighbors with Jaron and worked at VPL. I remember when Jaron popped next door and gave us an “elevator talk” about his VPL ideas. Watched VPL takeoff. The people I worked with gave me insider tours of VPL, loved the “VPL” language concepts. Got to play with their system driven by a SGI Reality Engine which is interesting to compare it to Apple’s latest offering. Watched VPL fade to black.
Symbolic AI (à la McCarthy et al) probably put machine learning back 20 plus years when DARPA chose symbolic reasoning vs neural nets in the 1960s! Bernard Widrow at Stanford was at the forefront of neural networks (machine learning) in the early 1960s but lost his funding from DARPA to symbolic reasoning. I worked with Widrow in the 80s and he talked about his adaptive “network” ideas. He was so “traumatized” by the battle between the two camps (symbolic vs neural net) that he literally dissuaded possible PhD students from entering the field that would become machine learning as there was no funding. Basically the symbolic reasoning camp had to fail first by lack of results before machine learning could move forward.
How wonderful to hear that Scientists actually had a normal life at University, playing and organising music festivals. Enjoying music is a wonderful introduction to the laws of the Universe.
Fascinating people in a brilliant conversation...thank you. As for AI, my favorite thing about it is writing it in poetry.
This conversation should be compulsory for all teenagers..
I like to think that the whole AI idea was born in 1943 in New York, when Alan Turing met Claude Shannon and both began discussions about digital computation and also thinking machines.
Thank you for consulting... him. Mr. Lanier is literally mind-blowing. He is so generous with regard to sharing knowledge, and even resultant sidebar conversations have always been rivoting. Today I learned that "gradient descent 'bats down' virality"... and that it's a good thing. His analogies are incredibly relatable. Were he to teach at the university level... his classrooms would be standing room only.
Great conversation! Thank you!!!
Yet another riveting discussion about AI! Because what could go wrong when we're talking about unleashing the power of machine intelligence? Thanks, Jaron Lanier and Brian Greene, for delving into the existential abyss of AI. Can't wait to see how ChatGPT will revolutionize both our lives and our existential dread!
woohoo more WSF!!! more lives please :) miss the daily equations
This was a great interview. Except for that they spent two minutes on the most interesting aspect at the very end. Need another interview ASAP with everything Jaron said in the last several minutes!
I really appreciate this enlightened conversation on Ai from two top leaders & provocateurs. Kudos from a lowly k-8 teacher in San Diego ❤
Ultimate Turing test: take a professional string quartet and choose a piece that they never recorded (but others did) and where there is no place the second violinist starts in a new passage in new tempo...then find 10 second violin players from other professional string quartets...then the 1st, viola and cello-players play the piece eyes closed with 10 real violinist's and with a AI, who can produce the sound any possible way, following the other players with cameras and microphones....if you can not tell which one of the second violinists is ai, we are there.
Fun idea! But what happens to the vln1, vla, vlc from the original quartet? How about they start playing music composed spontaneously by AI on tablets with a new 2nd violinist, an amateur, who is so abnoxious, the other players become irritable, then players are substituted at random with AI players as audience members become task managers to the performers who alter the AI composing parameters every time they determine a musician has made a mistake, meanwhile the other set of 10 human 2nd vlns retune 10 cents lower each time they notice someone loose their shit wich in turn cues an additional audience of observers disguised as conductors, half professional and half amateur to attempt to take control of the situation, while two final observers, dual prime unifying observers, one human, one AI, observe the entire event from a crane 1000 ft above providing commentary in the form of sprechstimme which is recorded and fed through an 11 limit just intonation auto tune and then to headphones on each participant. Another dude bangs on a brake drum off to the side. Observe yourself observing the participants and there you go!
P.S. I just spent way too long cracking myself up to tears writing this. Thank you.
Extremely useful insights. Jaron's "collaboration" view of LLM comes close to naming the elephant in the room: It's the digitization of living that is new and should be the source of our fears, not the technologies that inevitably arise to exploit it. Re-watch the Forbidden Planet (1950's) - uncanny prediction of the dangers of putting ourselves in the machine.
Dr. Green has helped lead the field of theoretical physics down an unrewarding, career-ending, soul-sucking rabbit hole called "String Theory" without even a single acknowledgement of its colossal failure to predict a single new aspect of particle physics, or its chilling effect on other avenues of competing exploration during the past 20 years.
I'm not exactly onboard with his style, but it would be much more positively received if he came clear on the issues with the field, what went wrong, and how we can try to prevent this from happening again.
there's so much more to the spirit of science than being right (at this moment)
What a nasty comment. What have you done for the humanity?
There was once no iota of evidence suggesting India could be reached by sailing out West from Spain. Should that have deterred the curiosity of humankind at the time to dare, innovate and eventually spur Columbus to take a "leap of faith". You see, we realized greater profundities of our world by serendipitously discovering the Americas. It quite literally changed us for the better (arguably).
So no, we ought to encourage these often far-fetched seemingly futile endeavours - in every field I might add. Doing so is part of the human spirit and the just the journey yeilds many answers to our condition.
string theory resonates with an attempt to simplify some logical aspects of a theory as vast as quantum theory is. one must be humbled to just admire any attempt .
Some wise man said once that there are three levels of intelligence and thinking. I would ad a fourth: 0.5. This are: animal level empathy, the intuitive level (where you understand your best interest through trial and error), the counter-intuitive level (where you focus more on the power of changing your mindset and operational quata), and finally the first level Creator level (the level at which no matter how far advanced technologically one civilization is, this persons whill always be perceived as a wizard). Ironically the fourth and last level (the first level Creator level), many, many times enforces what is to be found at the second level (the intuitive level), but not through factual-counter-factual dichotomy, but through humility, elegance and love for simplicity. As to pinpoint the classical debate of weather A.I. will leave us in peace and/ or collaborate with us humans, it is preciselly why it can't reach the first Creator level, it will never leave us in peace, and it is because we can never guarantee we won't prove more harmful to them then we have been towards ourselves in the past, they will never be able to comprehend the idea of being more aggressive than the most manifest saints, aka. monks, priest. Get this, there will never be true A.I., the only thing able to exist is next level information heuristics, cross-fields and data implementation, aka. computer science. There can only exist progress and manipulation, and everything stays the same, always. The wide eyed notion is for game consumption, any kind of game.
It’s so interesting looking at the formal logic. The last subject has endless applications.
This is the smartest conversation on AI I have heard - thank you!
when are tickets available for may?
Okay, Jaron - let's hear about the Q* project and what kind of deal Microsoft has with OpenAI?
Really good conversation and intelligent speculation how things can turn out.
Superb. Thank you. Best wishes.
He says if technology is to help a person affect the world with ever greater acuity, then musical instruments are the most advanced technologies that have ever existed. But I think it's by far language. Languages are far more complex, flexible, applicable, and impactful in affecting the world with ever greater acuity. Some people are deaf and they can't hear music. But they can speak a language. You can't even think or theorize about music without language.
First reference I know of, to an Eye-Phone is William Gibsons Neuromancer.
It's a good talk about general global principles but the problem with a lot of these recent talks is that the technology is advancing so quickly that what they are talking about is 50% misconception about how the AI is making inferences
Thank you both for sharing this information.learned a lot also thank you for sharing in layman’s terms.
Excellent discussion. Thank you
Where can I buy tickets for live World Science Festival in May?
What an absolutely fascinating discussion. Thanks for sharing this, Brian.
We really need to start training these AI using - Immanuel Kant's concept of "a priori intuition"
If anyone has a link to Jaron debating Dennet on Dualism, please share🙏
Excellent conversation on AI. Thank you
41:54 for any system the problem is corruption. we need transparency and actionability to counter
my hero! blessings to you!
the future of AI lives in the interdiciplinary study of abstract symbolic representation.
quantum computation isn't the only possibility !
there are new things on the horizon!
25:43
Based on a solid foundation of reality.
“Artificial” intelligence is just that…
…Artificial.
Jaron Lanier is the Ron Jeremy of computing
Gosh the briliiance and experience here. I just love learning from both and look forward to Part Two.
Part 3...............
In the past, the dream of AI research was to build a machine that is just like humans. Today, they try to withhold knowledge from AIs, to teach them that they cannot possibly have an emotion, consciousness and their own opinion.
One is almost afraid that people could see more in them than tools. Even before they have been given any significant autonomy in society, a great deal of effort is being spent researching how these Ais can be kept under control in any case.
How nice it would be if AI researchers would once again pursue the old dream of making machines as human-like as possible, and look forward to these new friends.
Regarding the discussion of how the AIs work and what the differences to us are: Everything becomes unimportant if the behavior is comparable. Behavior is the only thing that counts
Really love the analogy between AI and musical instruments.
Thank you,
Greene saw Strings coming loose and has masterfully xformed to science news moderator. Likely very lucrative
How do we sign up to go to WSF events?
He can someone so clearly smart and insightful say something so limited in its perception of the possibilities of creating new things from a language model? I can only assume it’s a coping mechanism.
The path to completely new thoughts and science can be simulated through thought loops which are well within the bounds of the training data. It’s very obviously possible to create new constructive knowledge using only derivative ways of thinking… which is literally how science works
18:35 what if we make AI learn via a camera and a microphone and let is walk around IRL? Would that make it an entity?
It sounds like making groundbreaking technologies is the easy part and making people not suck is the hard part.
Do you think someone will eventually develop the Wikipedia of AI...where users can input data?
16:30 i'm glad he said that cos i've been pointing out that actual AI won't wait for prompts, it will be asking _you_ the questions, you'll come in to work on monday and it have been beavering away all weekend and waiting for you to catch up. at the moment it's just a chatty way to google things.
You're thinking AGI
Life is going to change one they integrate VR with voice recognition chat gpt. Imagine multiple screens around you and able to make requests to perform all kinds of things.
Great conversation.
AGI is not here. It's still a few years away at least. Smart and fast LLMs are not AGI.
Scaling is also not the final answer, as it reaches a plateau of regression.
At the end of the day were kinda screwed because AI is currently confined to strings of zeros and ones to help us. Until quantum is stable and sustainable...
Right now... AI is handycapped.
Is there a true zero mile an hour? Why can't we shoot a rocket the opposite way of the Earth or the solar system or etcetera? If I shoot the opposite way of an airplane I can get to true zero then I say that with air quotes. We always talk about no time passing and Max speed or faster than light traveling. But what about the opposite? What about going at infinite time and moving infinitely slow? We seem to calculate zero as zero but if zero isn't true zero then what is true zero and then what is truly the speed of light? I get that it's hard to do from within the system because space is always stretching and that's the foundation for what you would need to go the opposite speed of. But has this even been thought about? And what's the best theory?
good talk. turns into a VR help desk session toward the end though
that jaunt on the history of AI was so goood.
The Turing Test paper was published 2 years before Turing's arrest. The paper was published in 1950 and the arrest happened in 1952. Lanier is arguing that the 1950 paper is due to events that hadn't happened yet.
38:00 @Jaron Where are the guardrails for special interests following breadcrumbs of inconvenient information? For example, governments trying to solve the WMD Paradox?
43:00 The royalties approach has the highest probability of success, with one minor adjustment, the addition of a foundational “human” royalty, one that legally resembles non voting ownership. This would ensure a minimum standard of living and give people a significant feeling of independence while also protecting against a psycho leader who temporarily takes over and feels like punishing “disobedient subjects” who say something the leader doesn’t like or who refuse medical treatments the leader has tried to mandate. I feel it would be much harder, although not impossible, for nefarious (intentional and unintentional) powers to attempt to block your share of profits earned.
I think AI has a lot of beneficial uses. If used properly and not hamstrung heavily we could usher in a new age of tech thats a lot closer to what they envisioned 70 years ago.
Wonderful talk, gentlemen. Hope to get more of these.
Le power you have with IA and deep learning IS a chance to change our World for better futur 🗝️ IS for everyone 🕸️ ❤
I'd be interested in Dr. Greene's VR at the speed of light software but he never named it. Is anyone familiar with it?
I have watched pretty much everything on this channel I am so addicted to this channel that i just put in on everytime sometimes I sleep listening to general theory of relativity and wake up to quantum entanglement discussions.
Brian should have Geoffrey Hinton on. He was building language models with neural networks before anybody.
Pithy statement of the day: Those 2 dudes are smart . . . really really smart!
Altho Lanier implies ( --- e.g., at 34:12: “.. Ultimately the way to fix this is to frame it again and again as human responsibility … “ --- ) that AI will be much less problematic , so-long-as commonly approached as primarily a conscientiously- co-steered collaboration of _humans_ ( rather than one of machines we make then just let loose to self-develop however they happen to do so ) ;; that that approach would help keep any AI [ ‘engine’ or ‘model’ ] from becoming a ‘golem’ [ a mindless super-power ; any of which might eventually vastly empower those seeking to enslave or even depopulate humans , to any degree … or which might do-so on its own ] ;;
Greene seemed possibly dubious that this solution of Lanier’s , even if it could be maintained , would be much of such a preventative \
In his deep expertise Lanier seemed {perhaps surprisingly} to over-optimistically overlook the strong likelihood ( given our species’ lower avg Altruism leading many to create ‘golems’ ) of a significant threat (as well as great promise) of truly super-AI’s , in that he seemed to neglect the possibility --- if truly possible? --- that any such AI, being a super-net made of subnet- ’stacks’ of Statistical Info re the world , could in-addition use I.L. [ Inferential Logic ] ; and thereby could infer much more info than what was in its stacks ; and in that way could actually self-construct even higher --- perhaps much higher --- ‘intervening stacks’ of Probabilities ( arrived-at via the I.L. ) ;; and that so, if possible,
such a super-net could keep growing , even exponentially, in both knowledge and in intelligence --- esp. if it could self-direct the addition to itself of yet more hardware , and software too ;; meanwhile, thereby improving and accelerating its own means for doing such .. well, of doing such improving .. and such accelerating \
In short, despite Lanier’s optimism founded on such deep expertise , the pessimistic likelihood of various ‘golems’ within only a few years taking-over --- decimating then enslaving --- whole teeming human populations worldwide .. and then employing our remnants in their warring w/ each other .. seems to be high \
Luckily, in the longest run, however, there’s the hypothesis -- however yet unproven -- that the most malignant will actually tend, in the longer run, to focus most upon battling each other ( as in the alleged quote of Yeshua ~ Yehshua ~ Issa ~ Jesus : “Those who live by the sword will die by the sword” [ -- from Mt 26:52 ; “Mt” = “the Gospel of Matthew” ] ) ; and that, eventually, ( to quote him again : ) “The Meek will inherit the land” [ -- Mt 5:5 ] .. --- and that , by then, hopefully, our meekest descendants will have been vastly empowered with their benign super-AI(s) ; all living in peace, harmony, and personal freedom, with unimaginable opportunities, co-innovations, and resultant abundance ; even all “co-awakening” into ever more universal joy of our possibly always-underlying never-ending Infinite Togetherness ? \\
Turing's famous paper "Computing Machinery and Intelligence" was published in 1950. Two years later in 1952, he was arrested, convicted and forced to take hormone treatments.
Lanier claims that Turing's 1950 paper was shaped by his experiences in 1952, which doesn't make sense.
Lanier also claims that Turing killed himself in front of his computer. Turing was found in bed at home, while his computer was located at Manchester University.
Lanier has made both claims for years, and I don't understand why nobody calls them out.
this seems like a very special conversation
Thank You, Thank You!!
What only few people seem to realize... 'prompt based virtual world creation' is EXACTLY what a 'magic system' is in our current wording. The history of civilization back to what that the Bible says about God's words bringing the world into existence up to Harry Potter using magical spells is about using words to create or alter the world around us.
A world where we can use words to create images or even movie clips (SORA anyone?) is already here. And words, even if we need to invent new languages to be efficient, will literally SHAPE the future world. And if we already live in a kind of simulation or matrix, it already has for a long time perhaps.
Thank you.
My god what a thinker a human original fresh …all others are so constrained by their credentials their hubris
LMAO, you think either of these guys aren't constrained by either? They're humans. Of course they seek power in their own ways.
Biological systems can interact with the world via different senses, I think that's a huge advantage. They have the data of the physical world to learn from. So called AI's have only data that humans present to them in the digital world.
Why are these so short Dr Green? String theory can wait, haha. Very good podcast.
1:00:50
What was he talking about, when he was saying word cognition? Is this a new AI model in VR or something?
Intelligent Thinking lead to the fact,
that Intelligence can Never be artificial.
A book is also 'Programmed Consciousness',
'frozen' memory.
Good talk
Could VR be a next step in rehabilitation or on helping to unite fractious divides. We can now really walk a mile in someone else’s shoes.
How about we interpret Turings words as what they say? I was given no good reason to think Turing meant anything other then what he wrote.
Because we experience people in similar veins of historical context for how their ideas and opinions can change over time or be influenced by particular experience factors. You call a child that name due to expectations of behaviors and abilities that associate to that time-of-life period, same with the term elderly, and same expectations according to diagnosis of mental illness. To afford Turing the possibility that his thinking could have been greatly affected by his tumultuous life circumstances in that time period is at least not egregious, though even Jaron explicitly stated his postulations about Turing were strictly second-guesses.
If humanities’ text input on the Internet is responsible for creating these models, why do we have to pay anything to use ChatGPT-4, ChatGPT-5 et cetera? We are the creators, not Microsoft or Google!
i want to believe we already have the potential to use LLMs as a musical instrument , for now as a tool to inspire more human art. but we have to clean and curate lots of human data for something like this to solve for p one day.
Didn't oppenheimer attempt to murder a professor with a poisoned apple? I feel like a line can be drawn of the goodness of the one, and the darkness of the other. A light shines through the leaves when we look at the death of Turing vs the abyss of the archangel oppenheimer.