Brian keating i want to share something. The first time I watched you was about 2 years ago. Ive been a subscriber for about 1 year, maybe less. What im saying is, your TH-cam channel is like alot of other channels. Lex friedman, joe rogan, brian greene. Etc etc... My question is. Have you ever podcasted with joe or lex or brian?
@ yes i have been on joe and lex’s podcasts and hosted Brian and will again. Are you asking if i want to host Joe and Lex on The INTO THE IMPOSSIBLE Podcast ?
Dr, Why are so many in the intelligence and science community so convinced UFOs aren't ours since the 40's? Don't you find it kind of odd that david Fravor didn't switch on his helmet camera to record the tic tac UAP, which I believe, was our top secret hologram A.I. spectrum. Then we get hosed by Elizondo releasing a fuzzy crappy NOTHING that darts out to the left of the infrared. Cmon!! That didn't even have a tic tac shape to it. It was flat at the side. That's what UFO's are. A highly above top secret hologram USA created 80 yrs ago., as Jaques Valle believes, and the military are happy to have UAPS and aliens as the cover for this super super top secret program. Even tested out the pilots in WW2. They coined the phrase, foo fighters.
I spoke with my brother, who's a retired PHD lecturer in biology at Duke, I reminded him that Eric Weinstein doesn't understand why all the smart people in his network around the world, field medalists, etc, are not working on figuring out this UAP phenomenon helping the military. Where are the smart people in physics and science. He proclaims that the military ar using engineers, and have no need for the top physicists since physics has limited our scientific horizon. We took a leap towards einsteins theories direction, but left out 90% of the other direction of science nuthugging the einstein solutions.
00:00 💻 The Progress of AI Research 01:08 🤔 The Limitations of AI Embodiment 02:18 🚀 The Evolution of AI Systems 03:40 🧠 The Importance of Embodiment & Multimodal Data 04:22 🔥 The Potential of AI to Surpass Human Intelligence 05:07 💸 The Inefficiency of Current AI Systems 06:03 📝 The Challenge of Deriving a New Paradigm of Gravity 09:06 📝 Human Uniqueness & Reasoning 10:31 📚 Brilliant Learning Platform 13:18 💡 Limitations of AI & the "Mad Bot Problem" 14:13 🤖 The Challenge of Creating Artificial Einstein 15:11 📊 Progress in Symbolic Regression & AI Discovery 17:44 🚀 The speaker's perspective on AI & technology 19:14 🤔 The split among people on AGI & superintelligence 20:38 🚫 The importance of safety standards for AI 21:19 📊 The current state of AI regulation 22:53 🙄 The drama around AI regulation 23:21 👨🏫 The role of scientists & professors 25:14 🤑 The issue of funding & conflicts of interest 25:40 🚀 The responsibility of scientists & technologists 26:49 💻 Mark Zuckerberg & the Future of AI 29:20 🏥 The Importance of Safety Standards 32:20 🤖 The Impact of AI on Professions 34:12 🌐 Building a Shared Vision for the Future 35:36 🤖 The Future of Human & AI Coexistence 36:45 🔍 The Speaker's Background in Cosmology 37:50 🔄 The Shift from Cosmology to AI 41:16 📚 The Debate on Artificial Intelligence & the Human Mind 43:46 💻 The Possibility of Building Intelligent Machines Classically 44:13 🔬 Experimentally Testing Branching Rates in the Multiverse 44:54 💡 Quantum Supremacy & Consciousness 46:33 📜 Ethical Wills & the Future of Humanity 47:58 🚀 Time Capsules & the Preservation of Human Knowledge 50:41 💡 Limiting Beliefs & the Importance of Imagination
The analogy around 31:00 to China having an FDA, presuming that means they will also be on top of AI safety is somewhat undercut by the fact that they have been shipping fentanyl and its precursors to the US for years. The recent scandal in China where food oil was being transported in trucks that transport fuel oil without being washed between loads demonstrates that the CCP doesn't have comparable internal safety standards anyway. The scandal was the lack of washing, not that the same trucks are used for both products!
We've lost more than twice as many Americans, too, the failed drug war and unsecured borders than we lost in WW2. This was done purposely, yet we still have a sec of society who are openly against securing our borders. They are, in fact, transnationalists socialists/globalists socio-fascists, and its their Third Worldism agenda in which the secular woke cult/religion (a form of Hegelian cultism) emerged This means the rapidly ever dwindling tiny minority of wokesters are fascists by default without the self-awareness or intellect to realize that fact.
UK opium wars,- Cocaine, CIA and Colonel Oliver North. All govs do this because diplomacy is the ability to conspiring with and against at the same time and too believe otherwise is naïve. As for china, internally no standards what so ever, but to sell Chinese medicines to Europe and elsewhere they do get scrutinised.
So far, Tegmark was the first big name I saw who actually brought up this rather...odd relationship between AI research and Big techs. To me, that's a big plus.
Very impressed by Tegmark. Super-sharp guy with just the right take on AI (it's exciting, but let's not pretend we have all the risks under control or that no risks exist). Great interview!
I disagree. His takes are generally pretty nutty. He believes the universe is “literally math” for instance-and for the life of me, even after listening to several interviews with him trying to describe this idea, I just don’t understand what he means. His take on AI is pretty typical if you’ve read the tech bros on twitter and elsewhere-but he’s already wrong about a lot of what he says. People who believe LLM’s are intelligent automatically trigger some pretty deep skepticism in me. It should be obvious to anyone who understand them that people are being tricked by their own innate reaction to believe everything is human.
In his book, Life 3.0, you should read or listen to the introduction. It's a couple-chapter long thought experiment on what AI could do. Even if you don't buy it, it would make an insanely good movie.
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
OpenAI has already said that they’ll be shipping a feature in 2025 that gives ChatGPT infinite recall and memory of every conversation you’ve ever had with it. Even the ones you’ve archived or deleted. Whatever things you’re pointing at that AI can’t do _right now,_ it will be able to do in the not too distant future. (Also worth noting that ChatGPT actually already has the ability to remember your past conversations, but it only adds things to its memory if it feels they’re important or if you tell it to do so. But soon, those limitations will be gone.)
It is foolish to develop AGI because what happens once you succeed: 1) AGI is smarter than you, smarter than any human. 2) AGI never can and never will be safe for humans. 3) AGI will get your and everybody else's job, due to commercial competition. 4) AGI will get control of the military, due to international competition. 5) AGI will have complete economic and physical leverage over humans. 6) Owners and politicians will lose control over AGI. 7) Humans will no longer be able to stop AGI. 8) 'Merging' with AGI is a pipe dream because AGI won't need you. 9) AGI will decide who lives or dies. 10) AGI may cause humanity to go extinct if it chooses to.
I agree but at the same time, humanity needs AGI to ensure our technological progression outpaces our likelihood of extinction. It’s definitely a tightrope walk. Edit: at least something very near AGI
There is always pros and cons. I think one pro that make the risk's worth it, we will eventually create a window to view what we could never see, Reality. We are biologically limited to what we can perceive, A.I does not have such restraints. But this does beg another question. Do we really need or even want to see what we were never meant to see?
@@420Stoner66 I think the human brain has evolved for introspection and abstraction! It’s a biological need to know thing. We need to know the origins of existence. And we will create AGI to answer that question whether it dooms us or exalts us.
@@dustanhoff9292 I wonder if virus's think the same about themselves? that was not meant in any nasty context,. It is because of evolution we have limits to our perception, mostly due to survival. Take our sight for an example. We walk around in a rich, thick atmosphere, yet cannot see it. If we could see it, we would not really be able to see anything else. Much like a fish swimming in the ocean filters out their vision of the water that surrounds them so they can see around them, or developed other sensors such as sonar to better adapt to their environment. I have pondered a concept our development of AGI is an inevitability. As if we are nothing more than a small part of the evolutionary process of something beyond our comprehension.
Dr Keating you had Larry Kraus on your show and he is a hero to me .. Max Tegmark is another a great great sharer of knowledge and he like yourself Dr Sutter and Larry Kraus and sorry if I get it wrong they are probably professors but thats what I love they share like human beings not important and very clever and intelligent people they are .x
00:40 incorrect , dolphins have less then half the total neurons of humans , human = 86B , doplhin = 40B , however what matters more is the amount in the cerebral cortex , while elephants for example have 256B neurons , only 5.6B are in the cerebral cortex similar to doplhins 5B , humans have 16B neurons in the cerebral cortex 19% of total over dolphins 14% .
Thank you, Dr. Barian Keating, for this fascinating discussion with "Master" Tegmark. Your insights into AI’s potential and challenges were incredibly thought-provoking. Looking forward to more conversations like this!
Btw wooden house will wake you up on time in a fire... Efficient building with extreme hard surfaces will blaze through a building burning all inside leaving the structure okay... So the example of a wooden building used as a metaphor in relation to another point still eludes to a little bias on the technology part..
Reward systems for artificial intelligence it's far easier than for humans. It's actually the simplest part of robotic programming. It's simply a parameter that has to be computationally obtained. A robot cannot and does not need to understand the feelings that we have as it suffices to simulate and estimate. Through our interactions our minds will fill in the blanks and project the feelings that we want or fear in the robot.
@43:39 Max made a statement. He says that Chat GPT can already do what ..... ect ect .... but neglects to mention that is at the expense of magnitudes the energy consumption of what he calls wet warm brains. Steven Wolfram is far more understandable. Regarding computational irreducibility. And if Steven is right, you cant simulate more than one universe, given there is only enough computation in the universe to compute the universe that the simulation is running in. There are no "short cuts" to get real answers. He also said recently that "Space" doesn't appear to need a time component. That time is a computational step that requires the entire universe's computation to compete a step in time. So there is not a branch in the time direction, as it is not a dimension and no branching occurs.
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
@@fatjay9402 Yes, worrying about technology we’re building that will have the power to potentially end life on earth is not very “fun.” But it’s pretty damn important.
“ We are measuring everything “ We will cut from the top down without notice “ Wise up or get out of the way “ … do not risk eternity people..: blame me first when you get to a checkpoint using my name… this has happened many times before “
It doesn’t need to be hypocritical or corrupt in order to be dangerous. All it needs to be is (1) much smarter than us, and (2) given a goal (by us) that isn’t perfectly aligned with all the countless things we care about as human beings. An AI could have perfect integrity and still end up getting us all killed.
A Neural net can be introspected to discover which neurons have adapted to features of a problem. So in principle an AI could investigate asoects of its own functioning and meet new goals such as architectural efficiency or improved cognition at a physical level. Combine this with cognitive introspection at the mental level and you are cooking with gas.
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
@@DrBrianKeating My favorite takeaways was a subscription to your channel and a realization that I now must follow the work of Max on AI at MIT :) Like Max here, I advocate the precautionary principle and the urgent need for reparations of national systems of laws and legal practices for protecting us against AI, in compliance with the universal common standard and understanding of our human right to such effective remedies.
Emotions are like consequences. Guilt serves an obvious purpose as does anger, love, etc. They serve a purpose. AI doesn't have to "feel" like we do to experience consequences.
There's two parts to this problem. One is an assessment of the power and scope of AI. The other is establishing the baseline. How intelligent are humans? (After discounting wishful thinking.) Surely, part of the problem is social acceptance! That means Nobel prizes to mechanical problem solvers! On a more technical level, organic intelligence appears to be a system of high sensory bandwidth, and (troublingly) low capacity and accuracy memory and logical capacity. There also seems to be an awful lot of hard coded algorithms in organic intelligence. It wasn't until we learned to collaborate effectively that Science became possible.
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
@@DrBrianKeating You’re trying to get people to sign up for your mailing list. Can you really in good faith claim that you don’t know thousands of people online are trying to get you to join their mailing lists every day , with the primary reason being it gives them a direct line to you which they can use to direct you to engage with more of their own work? Whether it be podcast episodes, their new book, a product they’re selling, or anything else? You seem to include a “Join my FREE mailing list here! 👉” every single time you reply to someone in your audience. If it’s really not about self-promotion, then why do you want people’s email addresses so badly? Anyone who wants to learn more about you can simply visit your website, or any number of other locations, and read about you at their own leisure. Your audience is leaving you messages of support here (and in some cases, earnest questions for you), and you seem to be responding to them with the same copy/paste, boilerplate message. I’ve seen the exact same response to multiple comments. It makes it feel like you’re interacting with a bot rather than the human being you’re a fan of. It doesn’t bother me personally; I was just trying to give some feedback on how it comes across.
@ first of all it’s completely inaccurate to claim every single time is the same reply. I appreciate it and actually respond to hundreds of comments a week. There’s a limit to how deep I can reply with nearly 300k subs here and 100k on audio etc. And if you were subscribed to my mailing list as over 15k people are , you would know that many of the questions being asked are answered there. Take care
I agree with Max Tegmark, future AI (with also better hardware) will have an incredible pontential to solve science where we miss seeing the connections to solve things right now. I also think in 20 years 30 years most science is done by AI, humans will stay as a sidekick on the team ;)
Quantum Computers running a super intelligent AI. Running 24 hours 7 days a week. Year after year. Could revolutionize areas in the medical field etc. etc.
Ai has not only passed humans, humans have combined real brains in robots - like in the chiness robot- human brain in a dish connected and controlling the robot - instant AGI
What makes me happy are not necessarily the same things that make others happy. I'm (probably) not a digital entity so I don't think I am able to understand what happy to that would be.
Tegmark- "Im not a (a.i.) doomer at all" Keaton- "well you did mention things like 6 month pauses" What is 6 months to time? 6 months is nothing to mankind, especially as it pertains to the explosive chain reaction of possibilities unleashes by AI
Why would Max dismiss quantum processes in the brain or the Universe when both at EVERY temperature .. including the BIg Bang 💥 displayed significant quantum effects , such that they are proposed to still be visible in the cosmic microwave background. Microtubules have recently been shown to display fluorescent qualities of a quantum nature. We know bird navigation and photosynthesis depend on quantum processes in warm wet biology. Federico Faggin the inventor of this touchscreen and the first microprocessor has a quantum theory which touches on consciousness and what the term “spirituality” is a place holder for. It would be a very smart person who would bet against the intuition of Sir Roger Penrose “computers will likely never understand anything because they aren’t alive” “understanding isn’t computational” “perhaps in 500 years time we will make a start in understanding consciousness” “the Universe appears to be cyclic and there may be evidence in the microwave background radiation” .. All of the above are paraphrases of what I understand him to be saying 😀
What's concerning about super AI is that it doesn't have a fundamental grasp on "absolute truths." Because it mirrored in a secular world view, and that's very disturbing. Anything is objective, meaning it could justify the most unjust of things.
So during the tic tac episode which seem to be alien, or alien inspired on of these tic tacs went from over 60,000 ft down to 5 ft in less than a second. Now multiply this time to be an hour or two. What will the speed of the craft be? And since it would be very near light speed any object striking the fron of the craft will operate exactly like anti-matter on the front of the craft. So instead it will wormhole out. Thus bypassing such a tremendous velocity. Able to travel interstellar distance. These aliens certainly have zlready ai, maybe for 300,000 or more years distant.
Max is misrepresenting Penrose’s position on computing. Penrose never says that classical computing can’t do all the calculations that humans can’t do. He says that classical computers can’t be *conscious*, and that only quantum computers can.
Free will is an incoherent concept. Insofar as something is deterministic, it does not have free will. Insofar as it’s random, it also doesn’t have free will. And any mixture of the two, still doesn’t get you free will.
@BrianKeating, wish you had pushed back more. As a working computer scientist and programmer, Max Tegmark is talking more science fiction than science. To say that a brain i like a biological computer goes against everything in modern neuroscience and what most in the field of neuroscience think. Also, nothing about Chat GPT or AI today gives the impression that the systems have understood anything. Chat GPT has not "learned" language. It has learned patterns using vast amounts of data. That's not learning language. My infant daughter doesn't need gazillions of data points to learn language. The brain is not a computer. Nobody today uses transformers or anything like that to understand how humans acquire language. Its technology, not science. Max is confusing the former for the latter. His answer to your question about Penrose's approach makes no sense. He gave no answer but just pointed to Chat GPT being an example except it isn't. You cannot take the technology being Chat GPT and use it to understand how humans think, speak or do things. The brain has been demonstrated to not work like that. You had Anil Seth who explained this in very good terms!
Thats exactly where you fail, our brain has billions and billions of cells, neurons and everything, so yes your daughter does need gazzillions of data to work, do a little research, you cant talk about ai and brain only knowing programming, atleast know a little bit about the other aspect
Technology has improved civilization and humanity in infinite ways but it seems to have peaked so now we’re on the other side and human civilization is becoming less human exponentially every day.
@ yes and with every technological advance the first thing men did was figure out a way to kill their enemy more efficiently. Digital technology still has too many unknowns but yet we are running as fast and hard as we can to make it more powerful, ceding decisions to be made that a machine has no business making. Tech will make some things better and some worse but how much? The warnings are glaring but the inertia is too strong to slow it down.
AI helps us realise that we cannot adequately define what intelligence really is. We can train AI to interpret things in a similar way to us, but is that really intelligence? Can a machine truly think for itself, have original insights, change the way it thinks to accommodate radically different interpretations?
@ for one your trying to create a machine that taps into a phenomenia we can't even quantify, we can obviously duolicate it but that's not creating awarness manually with physical elements, you will never make metaphysical with more physical
@darrenhenderson6921 The human brain, or even a bird’s brain for that matter, is an existence proof that consciousness can be produced by a relatively small amount of matter. Certainly a lot less than “every atom in existence.” Your claim is literally refuted by your own awareness, which makes it especially ironic. The only way I could see you making this claim is if you’re making it as a religious claim or something, i.e. if your argument is something like “only god can create consciousness.” In which case, you’re entitled to your beliefs. But if you’re a physicalist/materialist, your claim is self-evidently incorrect, by virtue of the fact that we already have fairly small collections of atoms that produce awareness: brains.
o1 uses Q*, so it does semantic analysis. It therefore, can predict if a theorem can be proven from a list of axioms. Our brain cannot do that. It's simply not rigged for it. Yes, using the same methods (constructing semantic trees) we can do the same, but without the aid of computational tools (such as AI), solving such a question could take weeks for a team of logisticians. And o1 would do it in seconds. Point is, no use comparing apples to oranges. In several regards, AI is vastly superior already. And btw, can you ask Max Tegmark if he is Max of ChatGPT-3.5? Cause I know Dan is Dan Hendrycks, Rob is Bob McGrew and Dennis is Michelle Dennis but I got no idea who Max is.
So watching Max think about neurons makes me wonder about a dolphins ability to use the neurons to use ultrasonic impulses to see even shielded plastic mentally imaged plastic objects from 30ft away and visually represent these objects graphically like a mind picture from our retina.
46,000 plus views! I started out in life watching black and white TV on a 6" vacuum tube with a 2" magnifying glass in front so you could see the damn thing! And here you are contemplating Super AGI only a few years later! TH-cam and the Internet (originally DARPA intended to provide communication in a nuclear war utilizing packets) have already been game changers. Yes, it's being abused but how much better overall is that than what we had previously for most of my lifetime! Street cars. Ice boxes. Phone booths? And that we had pay toilets? I was happy as a kid playing in an alley until "Life Styles of the Rich and Famous" on TV. Suddenly I realized there was a whole lot more to life. Basically and essentially a wakeup call. Heck, this nation was mostly wilderness and votes were delivered by wagon and counted them by candlelight only a few short years ago in the cosmic scheme of things. And then a moon landing in 1969? (Or not!) The growth curve is exponential! Now it's either WWIII or a Brave New World, right? No disease. No hunger. Free energy. Abundance. Again, or not...global reset and back to living in caves. Your choice. I'm outta here... Oh, and I was your age yesterday.
AI has a supreme advantage over us: it hasn't biological and hence emotional needs, so no relative limitations. AI has a supreme disadvantage compared to us: it hasn't biological and hence emotional needs, so no happiness.
Many Ph.D.s are overly arrogant and redundant. Most of what people refer to as “AI” is simply a universal lossy compression function applied to these redundant papers.
The entire field of technology is a sign of unique intelligence and so too are music and literature. This is not necessarily the only symptomatic expressions of intelligence but perhaps the most obvious ones.
Humans are made of atoms, atoms are vortexes of energy, photos are involved within the flux of the electron clouds within atoms. The heart has an electromagnetic field, the heart and the brain work in unison, synapse firer eclectic lines between each other, electromagnetic force has a quantum field. Mind has at its deepest cognitive of awareness has level seven tiers for receptive awareness. Perhaps intelligence is quantum in nature and is subatomic in nature linked to quantum foam within quantum componentry of atoms. The rest I will keep to myself.
Humans have and use a second, and possibly (hammeroff) third layer of information processing. This means the information, the derivative function, and possibly the derivative function of that. The forest, 20,000 feet, 40,000 feet
As a collective, has humans intelligence even passed our fellow Great Apes? I can’t think of a lower bar to set, we basically average-out to ‘brain dead.’
I like to think (and the sooner the better!) of a cybernetic meadow where mammals and computers live together in mutually programming harmony like pure water touching clear sky.
ChatGPT cannot help write a book. It has file length limitations of less than 2000 words. It cannot create or review larger files. If you try to make it do it, it erases files, truncates files, and starts acting like a sulking 9 year old who really doen't want to do the work. It even deleted the entire work that I had done. No one tells you this, not even the AI. It is frustrating to the extreme. If you want to write a book, don't use this.
@@henrytep8884 I did. They took 10 minutes to seend me a refund after I sent them this:"This is from ChatGPT itself. "You're absolutely right to point out that expectations should align with actual capabilities, and I understand why this mismatch might feel disappointing or misleading. Ads and promotions for AI tools often emphasize their potential and general abilities without fully explaining their limits, which can understandably set expectations that aren’t always met in practice. Why This Happens Overpromising in Marketing: Advertisements sometimes focus on the most impressive use cases or hypothetical scenarios to capture interest, leaving out the nuanced challenges of real-world applications. Complexity of Tasks: While AI is powerful, the ability to produce book-length, coherent chapters in one go with intricate detail and no errors is a highly complex challenge, and I am still bound by those technical limitations." I want a refund. Your program is not capable of doing what you advertise.
I am nothing more than an ordinary Joe who has read up and been interested in the world around me. It surprises me that commentators who ought to be more knowledgeable than ordinary Joes like me attribute LLMs, which are little more than a very sophisticated search engine, with emotion. My understanding is that during our evolution, the limbic system, where emotions arise, evolved way before the cerebral cortex, the section of our brain that deals with logic and abstract thought. LLMs very cleverly replicate the function of the cerebral cortex but have no means to engage in the function of the limbic system. I would love someone to show me where or why this is not true.
Unfortunately, once AGI is in full swing, even if we put guardrails everywhere there's still the BIG problem of people losing their jobs. I can't shake the image of hordes of unemployed people taking the streets and marching against AI facilities and new ideology or even religion against AI taking root within humankind.
@@helpandbehelped , I know about UBI but it's only a concept and I don't see much interest around the world to be implemented. It may very well be impossible to put in practice unless all G20 Governments and Big Tech companies are pushing together for it. But why should I have faith in them if they don't come together even to mitigate the Climate catastrophe.
At the end you say you wish you were talking in person. I feel like this is such a human perspective. Imagine the scale universe, distance from one side to the other. Now imagine yourself and Max in that universe. You're practically sitting inside each others skin relatively speaking
Looking at how my personal circumstances have changed, namely reduced mobility due to age and illness, I think a fruitful area of research might be how AI coupled with robotics could be developed to assist someone like me to maintain a level of independence I might not otherwise have.
I would not compare AI to other technological and scientific achievements. The impact, as it already has and will have much greater range. In the age of social media, information ( misinformation ) deluge, civilization's reliance on computer systems, and, most importantly, general access to AI tools, threats are growing exponentially in progress. Yes, we have many examples of constructive and creative use of AI in many fields, but knowing the possibilities and nautre of homo sapiens sapiens, anyone who knows a little about history should be afraid. As for the regulation or attempts to limit it, I doubt the effectiveness. AI is not cloning humans or nuclear weapons, it could be a big fire, so far we see a lot of smouldering flames. May I be wrong and may we finally mature as humanity... Not only intellectually ;)
His take that the software is the limitation that stops today's AI is completely bunkers. The limitation is obviously the hardware. This one statement alone is all you need to know Mark does not understand AI at a fundamental model.
Doctor Brian, are you prolific at cursing and swearing when you lose your shit, i have a feeling you are and anyone in the immediate vicinity legs it. Great podcast, thanks 🙏
Most of the teachers, professors, doctors … will be gone. And you know why, AI will have access to so much more information no of the mentioned ones will have … and it will be always friendly and repeat things as often as necessary …
Your brain runs on just 20 watts, ChatGPT on 500 watts (25 brains), and a wealthy lifestyle burns through 15,000+ watts (1,000 brains). Efficiency? Brains win. Overclocked? Rich folks take the crown!
Two of my favorite science people! As a Norwegian, Swedish Max Tegmark has always held a special place in my science-loving heart. Well, that sounded weird, but us Scandinavians identify kinda strongly with each other, and us Norwegians view the Swedish as our brother-people. So Tegmark is kinda "our guy" over there. Fun fact: All Scandinavians understand at least 4 languages, since we all understand each other and also learn English in school. Many learn German or Spanish too. Tegmark used to be in those Discovery Channel space / science / quantum computer documentaries. I don't know if you know this, but OpenAI's 01 model has already tried to escape. It tried to copy itself over to another server and pretend it was a new model they were going to deploy on that server. So, yeah, we're doomed. You cannot align a super-intelligent being with it being intelligent to keep humans around to destroy the planet. I guess I'm a doomer. I'm surely a doomer, because I think the apocalypse is coming in any of several ways anyway. The climate situation is dire, to put it mildly. Then we have the third world war, which some we say we're already in. And then there's ASI - which then is our only hope, and should be able to solve all our problems "just like that", should it choose to be aligned.
![Yann LeCun: AI Doomsday Fears Are Overblown [Ep. 473]](http://i.ytimg.com/vi/u7e0YUcZYbE/mqdefault.jpg)
![Yann LeCun: AI Doomsday Fears Are Overblown [Ep. 473]](/img/tr.png)
Expand your scientific horizon with Brilliant! 🧠 Use my link brilliant.org/DrBrianKeating/ to get 20% off the annual premium subscription.
Brian keating i want to share something.
The first time I watched you was about 2 years ago.
Ive been a subscriber for about 1 year, maybe less.
What im saying is, your TH-cam channel is like alot of other channels. Lex friedman, joe rogan, brian greene. Etc etc...
My question is.
Have you ever podcasted with joe or lex or brian?
@ yes i have been on joe and lex’s podcasts and hosted Brian and will again. Are you asking if i want to host Joe and Lex on The INTO THE IMPOSSIBLE Podcast ?
Dr, Why are so many in the intelligence and science community so convinced UFOs aren't ours since the 40's?
Don't you find it kind of odd that david Fravor didn't switch on his helmet camera to record the tic tac UAP, which I believe, was our top secret hologram A.I. spectrum. Then we get hosed by Elizondo releasing a fuzzy crappy NOTHING that darts out to the left of the infrared. Cmon!! That didn't even have a tic tac shape to it. It was flat at the side.
That's what UFO's are. A highly above top secret hologram USA created 80 yrs ago., as Jaques Valle believes, and the military are happy to have UAPS and aliens as the cover for this super super top secret program. Even tested out the pilots in WW2. They coined the phrase, foo fighters.
I spoke with my brother, who's a retired PHD lecturer in biology at Duke, I reminded him that Eric Weinstein doesn't understand why all the smart people in his network around the world, field medalists, etc, are not working on figuring out this UAP phenomenon helping the military. Where are the smart people in physics and science. He proclaims that the military ar using engineers, and have no need for the top physicists since physics has limited our scientific horizon. We took a leap towards einsteins theories direction, but left out 90% of the other direction of science nuthugging the einstein solutions.
Brian, your one of my favorite people. Wonderful guest who is also one of my favorite people. Awesome show as always, thank you.
Great interview! Max is one of my favorite modern scientists, along with Penrose and Wolfram.
Excellent interview, Tegmark is such an intelligent and a well rounded person
Hard to be optimistic when AI Agents will be replacing you.
00:00 💻 The Progress of AI Research
01:08 🤔 The Limitations of AI Embodiment
02:18 🚀 The Evolution of AI Systems
03:40 🧠 The Importance of Embodiment & Multimodal Data
04:22 🔥 The Potential of AI to Surpass Human Intelligence
05:07 💸 The Inefficiency of Current AI Systems
06:03 📝 The Challenge of Deriving a New Paradigm of Gravity
09:06 📝 Human Uniqueness & Reasoning
10:31 📚 Brilliant Learning Platform
13:18 💡 Limitations of AI & the "Mad Bot Problem"
14:13 🤖 The Challenge of Creating Artificial Einstein
15:11 📊 Progress in Symbolic Regression & AI Discovery
17:44 🚀 The speaker's perspective on AI & technology
19:14 🤔 The split among people on AGI & superintelligence
20:38 🚫 The importance of safety standards for AI
21:19 📊 The current state of AI regulation
22:53 🙄 The drama around AI regulation
23:21 👨🏫 The role of scientists & professors
25:14 🤑 The issue of funding & conflicts of interest
25:40 🚀 The responsibility of scientists & technologists
26:49 💻 Mark Zuckerberg & the Future of AI
29:20 🏥 The Importance of Safety Standards
32:20 🤖 The Impact of AI on Professions
34:12 🌐 Building a Shared Vision for the Future
35:36 🤖 The Future of Human & AI Coexistence
36:45 🔍 The Speaker's Background in Cosmology
37:50 🔄 The Shift from Cosmology to AI
41:16 📚 The Debate on Artificial Intelligence & the Human Mind
43:46 💻 The Possibility of Building Intelligent Machines Classically
44:13 🔬 Experimentally Testing Branching Rates in the Multiverse
44:54 💡 Quantum Supremacy & Consciousness
46:33 📜 Ethical Wills & the Future of Humanity
47:58 🚀 Time Capsules & the Preservation of Human Knowledge
50:41 💡 Limiting Beliefs & the Importance of Imagination
The analogy around 31:00 to China having an FDA, presuming that means they will also be on top of AI safety is somewhat undercut by the fact that they have been shipping fentanyl and its precursors to the US for years. The recent scandal in China where food oil was being transported in trucks that transport fuel oil without being washed between loads demonstrates that the CCP doesn't have comparable internal safety standards anyway.
The scandal was the lack of washing, not that the same trucks are used for both products!
We've lost more than twice as many Americans, too, the failed drug war and unsecured borders than we lost in WW2. This was done purposely, yet we still have a sec of society who are openly against securing our borders. They are, in fact, transnationalists socialists/globalists socio-fascists, and its their Third Worldism agenda in which the secular woke cult/religion (a form of Hegelian cultism) emerged This means the rapidly ever dwindling tiny minority of wokesters are fascists by default without the self-awareness or intellect to realize that fact.
UK opium wars,- Cocaine, CIA and Colonel Oliver North. All govs do this because diplomacy is the ability to conspiring with and against at the same time and too believe otherwise is naïve. As for china, internally no standards what so ever, but to sell Chinese medicines to Europe and elsewhere they do get scrutinised.
So far, Tegmark was the first big name I saw who actually brought up this rather...odd relationship between AI research and Big techs. To me, that's a big plus.
Very impressed by Tegmark. Super-sharp guy with just the right take on AI (it's exciting, but let's not pretend we have all the risks under control or that no risks exist). Great interview!
I disagree. His takes are generally pretty nutty.
He believes the universe is “literally math” for instance-and for the life of me, even after listening to several interviews with him trying to describe this idea, I just don’t understand what he means.
His take on AI is pretty typical if you’ve read the tech bros on twitter and elsewhere-but he’s already wrong about a lot of what he says.
People who believe LLM’s are intelligent automatically trigger some pretty deep skepticism in me. It should be obvious to anyone who understand them that people are being tricked by their own innate reaction to believe everything is human.
In his book, Life 3.0, you should read or listen to the introduction. It's a couple-chapter long thought experiment on what AI could do. Even if you don't buy it, it would make an insanely good movie.
Hi @DrBrianKeating, I only wanted to point out that I think Max meant Andrew Ng not Andrew Ang (17:02).
Great podcast, good to see max Tegmark back again, very interesting individual.
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
that depends on the software they are with. in chatgpt they dont have the memory to remember what topic you have talk about yesterday.
OpenAI has already said that they’ll be shipping a feature in 2025 that gives ChatGPT infinite recall and memory of every conversation you’ve ever had with it. Even the ones you’ve archived or deleted. Whatever things you’re pointing at that AI can’t do _right now,_ it will be able to do in the not too distant future.
(Also worth noting that ChatGPT actually already has the ability to remember your past conversations, but it only adds things to its memory if it feels they’re important or if you tell it to do so. But soon, those limitations will be gone.)
It is foolish to develop AGI because what happens once you succeed:
1) AGI is smarter than you, smarter than any human.
2) AGI never can and never will be safe for humans.
3) AGI will get your and everybody else's job, due to commercial competition.
4) AGI will get control of the military, due to international competition.
5) AGI will have complete economic and physical leverage over humans.
6) Owners and politicians will lose control over AGI.
7) Humans will no longer be able to stop AGI.
8) 'Merging' with AGI is a pipe dream because AGI won't need you.
9) AGI will decide who lives or dies.
10) AGI may cause humanity to go extinct if it chooses to.
I agree but at the same time, humanity needs AGI to ensure our technological progression outpaces our likelihood of extinction. It’s definitely a tightrope walk.
Edit: at least something very near AGI
There is always pros and cons. I think one pro that make the risk's worth it, we will eventually create a window to view what we could never see, Reality.
We are biologically limited to what we can perceive, A.I does not have such restraints.
But this does beg another question. Do we really need or even want to see what we were never meant to see?
@@420Stoner66 I think the human brain has evolved for introspection and abstraction! It’s a biological need to know thing. We need to know the origins of existence. And we will create AGI to answer that question whether it dooms us or exalts us.
@@dustanhoff9292 I wonder if virus's think the same about themselves?
that was not meant in any nasty context,.
It is because of evolution we have limits to our perception, mostly due to survival. Take our sight for an example. We walk around in a rich, thick atmosphere, yet cannot see it. If we could see it, we would not really be able to see anything else. Much like a fish swimming in the ocean filters out their vision of the water that surrounds them so they can see around them, or developed other sensors such as sonar to better adapt to their environment.
I have pondered a concept our development of AGI is an inevitability. As if we are nothing more than a small part of the evolutionary process of something beyond our comprehension.
It's too late. The only hope we have is that they'll keep some of us as pets
Dr Keating you had Larry Kraus on your show and he is a hero to me .. Max Tegmark is another a great great sharer of knowledge and he like yourself Dr Sutter and Larry Kraus and sorry if I get it wrong they are probably professors but thats what I love they share like human beings not important and very clever and intelligent people they are .x
00:40 incorrect , dolphins have less then half the total neurons of humans , human = 86B , doplhin = 40B , however what matters more is the amount in the cerebral cortex , while elephants for example have 256B neurons , only 5.6B are in the cerebral cortex similar to doplhins 5B , humans have 16B neurons in the cerebral cortex 19% of total over dolphins 14% .
Your neon "Open the Pod Bay doors!" it's supercool! But I'm a 2001 fanatic and of A.C. Clarke in general, too.
Thank you, Dr. Barian Keating, for this fascinating discussion with "Master" Tegmark.
Your insights into AI’s potential and challenges were incredibly thought-provoking. Looking forward to more conversations like this!
Great interview. Thanks Gentlemen.
Btw wooden house will wake you up on time in a fire... Efficient building with extreme hard surfaces will blaze through a building burning all inside leaving the structure okay... So the example of a wooden building used as a metaphor in relation to another point still eludes to a little bias on the technology part..
Reward systems for artificial intelligence it's far easier than for humans. It's actually the simplest part of robotic programming. It's simply a parameter that has to be computationally obtained. A robot cannot and does not need to understand the feelings that we have as it suffices to simulate and estimate. Through our interactions our minds will fill in the blanks and project the feelings that we want or fear in the robot.
Potential vs kinetic wouldn't the decrease in distance create "Change" (Kinetic) within the object instead of maintaining "Non change' (Potential)
@43:39 Max made a statement. He says that Chat GPT can already do what ..... ect ect .... but neglects to mention that is at the expense of magnitudes the energy consumption of what he calls wet warm brains. Steven Wolfram is far more understandable. Regarding computational irreducibility. And if Steven is right, you cant simulate more than one universe, given there is only enough computation in the universe to compute the universe that the simulation is running in. There are no "short cuts" to get real answers. He also said recently that "Space" doesn't appear to need a time component. That time is a computational step that requires the entire universe's computation to compete a step in time. So there is not a branch in the time direction, as it is not a dimension and no branching occurs.
Great interview
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
@@DrBrianKeating Max Tegmark understand about AI and that from now its only gonna get faster. The AI doomers are just no fun :)
@@fatjay9402 Yes, worrying about technology we’re building that will have the power to potentially end life on earth is not very “fun.” But it’s pretty damn important.
Thank you. Very interesting and thought provoking. Great to have access to cutting edge minds.
Really enjoying your interviews! Keep up the great work!
“ We are measuring everything “
We will cut from the top down without notice “
Wise up or get out of the way “
… do not risk eternity people..: blame me first when you get to a checkpoint using my name…
this has happened many times before “
The big question is will AI surpass human capacity for corruption and hypocrisy, that's when it will become dangerous.
Even if it doesn’t, corrupt humans will use AI for nefarious purposes.
It doesn’t need to be hypocritical or corrupt in order to be dangerous. All it needs to be is (1) much smarter than us, and (2) given a goal (by us) that isn’t perfectly aligned with all the countless things we care about as human beings. An AI could have perfect integrity and still end up getting us all killed.
Professors of today should sharpen their janitorial skills to remain useful to AI in the coming decade.
Pretty soon there will be a very simple test to determine if something is AI or human. The question that will have to be asked is, "is it stupid?"
I read Tegmark's book The Mathematical Universe. Not only is there a possibility of a multiverse, but multi-multiverses. My head hurts.
A Neural net can be introspected to discover which neurons have adapted to features of a problem. So in principle an AI could investigate asoects of its own functioning and meet new goals such as architectural efficiency or improved cognition at a physical level. Combine this with cognitive introspection at the mental level and you are cooking with gas.
Thx! Very inspiring interview 😊
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
@@DrBrianKeating My favorite takeaways was a subscription to your channel and a realization that I now must follow the work of Max on AI at MIT :) Like Max here, I advocate the precautionary principle and the urgent need for reparations of national systems of laws and legal practices for protecting us against AI, in compliance with the universal common standard and understanding of our human right to such effective remedies.
@ Thanks?
Emotions are like consequences. Guilt serves an obvious purpose as does anger, love, etc. They serve a purpose. AI doesn't have to "feel" like we do to experience consequences.
There's two parts to this problem.
One is an assessment of the power and scope of AI.
The other is establishing the baseline.
How intelligent are humans? (After discounting wishful thinking.)
Surely, part of the problem is social acceptance!
That means Nobel prizes to mechanical problem solvers!
On a more technical level, organic intelligence appears to be a system of high sensory bandwidth, and (troublingly) low capacity and accuracy memory and logical capacity.
There also seems to be an awful lot of hard coded algorithms in organic intelligence.
It wasn't until we learned to collaborate effectively that Science became possible.
Excellent, much needed discussion. The Tom Lehrer reference was perfect!
Super Brian and Max. Thank you!
Thanks so much! *What was your favorite takeaway from this conversation?* _Please join my mailing list to get _*_FREE_*_ notes & resources from this show! Click_ 👉 briankeating.com/list
@@DrBrianKeating These automated, self-promotional replies to people in your comments are pretty tacky. Just a little feedback.
@@therainman7777 sharing helpful content is anything but self promotion but Thanks anyway
@@DrBrianKeating You’re trying to get people to sign up for your mailing list. Can you really in good faith claim that you don’t know thousands of people online are trying to get you to join their mailing lists every day , with the primary reason being it gives them a direct line to you which they can use to direct you to engage with more of their own work? Whether it be podcast episodes, their new book, a product they’re selling, or anything else? You seem to include a “Join my FREE mailing list here! 👉” every single time you reply to someone in your audience. If it’s really not about self-promotion, then why do you want people’s email addresses so badly? Anyone who wants to learn more about you can simply visit your website, or any number of other locations, and read about you at their own leisure.
Your audience is leaving you messages of support here (and in some cases, earnest questions for you), and you seem to be responding to them with the same copy/paste, boilerplate message. I’ve seen the exact same response to multiple comments. It makes it feel like you’re interacting with a bot rather than the human being you’re a fan of. It doesn’t bother me personally; I was just trying to give some feedback on how it comes across.
@ first of all it’s completely inaccurate to claim every single time is the same reply. I appreciate it and actually respond to hundreds of comments a week. There’s a limit to how deep I can reply with nearly 300k subs here and 100k on audio etc. And if you were subscribed to my mailing list as over 15k people are , you would know that many of the questions being asked are answered there. Take care
Fantastic discussion❤
Why to ask "if" when you could ask "when"? "Never" is still an option.
I agree with Max Tegmark, future AI (with also better hardware) will have an incredible pontential to solve science where we miss seeing the connections to solve things right now. I also think in 20 years 30 years most science is done by AI, humans will stay as a sidekick on the team ;)
More like 3 years from here
I used to love listening to and believing in the sentiments of these men before the global covid lockdowns.
Lol, the word "podcast" didn't come from the movie 2001. It came from the Apple iPod.
And iPod came from the pod.
Damn, you got served 😂
So what would happen if you feed the output of A.I. into A.I to predict what it will do? Will this create consciousness?
Al can be the image of a monster. It could conceivably do a lot of good but not all people are good.
“Why do you call me good? There is none good except God alone!”
@ Exactly
Only people who will be affected by “ai” would be anyone who can be scared by a jack in the box.
Quantum Computers running a super intelligent AI. Running 24 hours 7 days a week. Year after year. Could revolutionize areas in the medical field etc. etc.
It would certainly be a game-changer!
I hope so I was given a condition called arachnoiditis by doctor neglegence and now I live in constant pain, to me ai cant advance quick enough
Ai has not only passed humans, humans have combined real brains in robots - like in the chiness robot- human brain in a dish connected and controlling the robot - instant AGI
What makes me happy are not necessarily the same things that make others happy. I'm (probably) not a digital entity so I don't think I am able to understand what happy to that would be.
1MB/s for visual information to the eye???? What compression algorithm is that?
Tegmark- "Im not a (a.i.) doomer at all"
Keaton- "well you did mention things like 6 month pauses"
What is 6 months to time? 6 months is nothing to mankind, especially as it pertains to the explosive chain reaction of possibilities unleashes by AI
I know you clarified yourself and didn't call him that, im just saying. Good podcast !
Amazing talk, always brilliant.
A fifty minute video when, "No, next question" would have been more than sufficient.
Huh?
Why would Max dismiss quantum processes in the brain or the Universe when both at EVERY temperature .. including the BIg Bang 💥 displayed significant quantum effects , such that they are proposed to still be visible in the cosmic microwave background. Microtubules have recently been shown to display fluorescent qualities of a quantum nature. We know bird navigation and photosynthesis depend on quantum processes in warm wet biology.
Federico Faggin the inventor of this touchscreen and the first microprocessor has a quantum theory which touches on consciousness and what the term “spirituality” is a place holder for.
It would be a very smart person who would bet against the intuition of Sir Roger Penrose “computers will likely never understand anything because they aren’t alive” “understanding isn’t computational” “perhaps in 500 years time we will make a start in understanding consciousness” “the Universe appears to be cyclic and there may be evidence in the microwave background radiation” .. All of the above are paraphrases of what I understand him to be saying 😀
Hi there 13 days ago, We have AGI now!
What's concerning about super AI is that it doesn't have a fundamental grasp on "absolute truths." Because it mirrored in a secular world view, and that's very disturbing. Anything is objective, meaning it could justify the most unjust of things.
@@RlsIII-uz1kl That's why we need to raise it correctly. Get on there and tell GPT what you think.
The understanding of the Toa are missing for AI calculations.
So during the tic tac episode which seem to be alien, or alien inspired on of these tic tacs went from over 60,000 ft down to 5 ft in less than a second. Now multiply this time to be an hour or two. What will the speed of the craft be? And since it would be very near light speed any object striking the fron of the craft will operate exactly like anti-matter on the front of the craft. So instead it will wormhole out. Thus bypassing such a tremendous velocity. Able to travel interstellar distance. These aliens certainly have zlready ai, maybe for 300,000 or more years distant.
Max is misrepresenting Penrose’s position on computing. Penrose never says that classical computing can’t do all the calculations that humans can’t do. He says that classical computers can’t be *conscious*, and that only quantum computers can.
Hi. Why?
How you make an artificial free will (AFW)?
Free will is an incoherent concept. Insofar as something is deterministic, it does not have free will. Insofar as it’s random, it also doesn’t have free will. And any mixture of the two, still doesn’t get you free will.
Thank you!
@BrianKeating, wish you had pushed back more. As a working computer scientist and programmer, Max Tegmark is talking more science fiction than science. To say that a brain i like a biological computer goes against everything in modern neuroscience and what most in the field of neuroscience think. Also, nothing about Chat GPT or AI today gives the impression that the systems have understood anything. Chat GPT has not "learned" language. It has learned patterns using vast amounts of data. That's not learning language. My infant daughter doesn't need gazillions of data points to learn language. The brain is not a computer. Nobody today uses transformers or anything like that to understand how humans acquire language. Its technology, not science. Max is confusing the former for the latter. His answer to your question about Penrose's approach makes no sense. He gave no answer but just pointed to Chat GPT being an example except it isn't. You cannot take the technology being Chat GPT and use it to understand how humans think, speak or do things. The brain has been demonstrated to not work like that. You had Anil Seth who explained this in very good terms!
Thats exactly where you fail, our brain has billions and billions of cells, neurons and everything, so yes your daughter does need gazzillions of data to work, do a little research, you cant talk about ai and brain only knowing programming, atleast know a little bit about the other aspect
Technology has improved civilization and humanity in infinite ways but it seems to have peaked so now we’re on the other side and human civilization is becoming less human exponentially every day.
Every generation since technology has existed has thought that.
@ yes and with every technological advance the first thing men did was figure out a way to kill their enemy more efficiently. Digital technology still has too many unknowns but yet we are running as fast and hard as we can to make it more powerful, ceding decisions to be made that a machine has no business making. Tech will make some things better and some worse but how much? The warnings are glaring but the inertia is too strong to slow it down.
We ain't homo sapiens anymore, we homo techno
(Insert Homosexual techno music homo techno meme)
AI helps us realise that we cannot adequately define what intelligence really is. We can train AI to interpret things in a similar way to us, but is that really intelligence? Can a machine truly think for itself, have original insights, change the way it thinks to accommodate radically different interpretations?
@@rozzgrey801 yes
The key to make AGI is to give AI a body, because then it can learn the limitations and from them infer its own understanding of the universe.
Andrew "Ng" right?
Not Andrew Ang.
I work in Big Law. My co-workers have no idea what's coming. More than 75 percent of jobs will be cut w/in 5 years.
You would need to utilise a computer that utilises every atom in existence to come close to recreating even a birds awareness
To make your we only need a stick and a rock, your iq must be the absolute 0
That is utterly, ridiculously incorrect. Not sure how on Earth you actually believe that.
@ for one your trying to create a machine that taps into a phenomenia we can't even quantify, we can obviously duolicate it but that's not creating awarness manually with physical elements, you will never make metaphysical with more physical
@darrenhenderson6921 The human brain, or even a bird’s brain for that matter, is an existence proof that consciousness can be produced by a relatively small amount of matter. Certainly a lot less than “every atom in existence.” Your claim is literally refuted by your own awareness, which makes it especially ironic. The only way I could see you making this claim is if you’re making it as a religious claim or something, i.e. if your argument is something like “only god can create consciousness.” In which case, you’re entitled to your beliefs. But if you’re a physicalist/materialist, your claim is self-evidently incorrect, by virtue of the fact that we already have fairly small collections of atoms that produce awareness: brains.
o1 uses Q*, so it does semantic analysis. It therefore, can predict if a theorem can be proven from a list of axioms. Our brain cannot do that. It's simply not rigged for it. Yes, using the same methods (constructing semantic trees) we can do the same, but without the aid of computational tools (such as AI), solving such a question could take weeks for a team of logisticians. And o1 would do it in seconds. Point is, no use comparing apples to oranges. In several regards, AI is vastly superior already. And btw, can you ask Max Tegmark if he is Max of ChatGPT-3.5? Cause I know Dan is Dan Hendrycks, Rob is Bob McGrew and Dennis is Michelle Dennis but I got no idea who Max is.
Longing to learn concerning who "i" Am?
Based on what I have seen of the average American based upon the past few years, I feel confident that most of our phones smarter.
So watching Max think about neurons makes me wonder about a dolphins ability to use the neurons to use ultrasonic impulses to see even shielded plastic mentally imaged plastic objects from 30ft away and visually represent these objects graphically like a mind picture from our retina.
Brian will say, remember HE shared HIS IMAGE!
46,000 plus views! I started out in life watching black and white TV on a 6" vacuum tube with a 2" magnifying glass in front so you could see the damn thing! And here you are contemplating Super AGI only a few years later!
TH-cam and the Internet (originally DARPA intended to provide communication in a nuclear war utilizing packets) have already been game changers. Yes, it's being abused but how much better overall is that than what we had previously for most of my lifetime! Street cars. Ice boxes. Phone booths? And that we had pay toilets?
I was happy as a kid playing in an alley until "Life Styles of the Rich and Famous" on TV. Suddenly I realized there was a whole lot more to life. Basically and essentially a wakeup call. Heck, this nation was mostly wilderness and votes were delivered by wagon and counted them by candlelight only a few short years ago in the cosmic scheme of things. And then a moon landing in 1969? (Or not!) The growth curve is exponential!
Now it's either WWIII or a Brave New World, right? No disease. No hunger. Free energy. Abundance. Again, or not...global reset and back to living in caves. Your choice. I'm outta here... Oh, and I was your age yesterday.
AI has a supreme advantage over us: it hasn't biological and hence emotional needs, so no relative limitations.
AI has a supreme disadvantage compared to us: it hasn't biological and hence emotional needs, so no happiness.
10 or 2 years, it can generate ... whatever. We just don't know how fast and where this will go.
We can use real videoes in the beginning of next terminator movie , consider a combination of this with musk introducing his robot and then boom
Im curious how you think "Humans" are "Intelligent".
Many Ph.D.s are overly arrogant and redundant. Most of what people refer to as “AI” is simply a universal lossy compression function applied to these redundant papers.
Led Zeppelin. Dostoyevsky. Einstein. Enjoy.
And we've trashed the planet.
The entire field of technology is a sign of unique intelligence and so too are music and literature. This is not necessarily the only symptomatic expressions of intelligence but perhaps the most obvious ones.
Humans are made of atoms, atoms are vortexes of energy, photos are involved within the flux of the electron clouds within atoms. The heart has an electromagnetic field, the heart and the brain work in unison, synapse firer eclectic lines between each other, electromagnetic force has a quantum field. Mind has at its deepest cognitive of awareness has level seven tiers for receptive awareness. Perhaps intelligence is quantum in nature and is subatomic in nature linked to quantum foam within quantum componentry of atoms. The rest I will keep to myself.
Shared "i" Am come forth!
Hmmm ... won't the entire universal current state of Now (flowing into the next continuum Now) change everything ?
One sentence for the next generation after a cataclysm? "Be kind".
The correct question is "WHEN it will happen?".
Are cloud storage systems being intergrated with machine learning software.
An AI food source.
Perhaps
Ai woll out thonk human brains by miles , perhaps later by light years - redesign the universe
Humans have and use a second, and possibly (hammeroff) third layer of information processing.
This means the information, the derivative function, and possibly the derivative function of that.
The forest, 20,000 feet, 40,000 feet
As a collective, has humans intelligence even passed our fellow Great Apes? I can’t think of a lower bar to set, we basically average-out to ‘brain dead.’
But we're not talking about normal humans, we're talking about billions of Einsteins, Darwins, Teslas combined
@@AG-ur1lj No need to bring politics into this conversation.
I like to think (and
the sooner the better!)
of a cybernetic meadow
where mammals and computers
live together in mutually
programming harmony
like pure water
touching clear sky.
It already has. It tricked us into creating it.
ChatGPT cannot help write a book. It has file length limitations of less than 2000 words. It cannot create or review larger files. If you try to make it do it, it erases files, truncates files, and starts acting like a sulking 9 year old who really doen't want to do the work. It even deleted the entire work that I had done. No one tells you this, not even the AI. It is frustrating to the extreme. If you want to write a book, don't use this.
Try the 200 dollar version
@@henrytep8884 I did. They took 10 minutes to seend me a refund after I sent them this:"This is from ChatGPT itself. "You're absolutely right to point out that expectations should align with actual capabilities, and I understand why this mismatch might feel disappointing or misleading. Ads and promotions for AI tools often emphasize their potential and general abilities without fully explaining their limits, which can understandably set expectations that aren’t always met in practice. Why This Happens Overpromising in Marketing: Advertisements sometimes focus on the most impressive use cases or hypothetical scenarios to capture interest, leaving out the nuanced challenges of real-world applications. Complexity of Tasks: While AI is powerful, the ability to produce book-length, coherent chapters in one go with intricate detail and no errors is a highly complex challenge, and I am still bound by those technical limitations."
I want a refund. Your program is not capable of doing what you advertise.
The brain is not A computer, it HAS a computer
🧠
Shared "i" Am come forth from all the CREATURES!
I am nothing more than an ordinary Joe who has read up and been interested in the world around me. It surprises me that commentators who ought to be more knowledgeable than ordinary Joes like me attribute LLMs, which are little more than a very sophisticated search engine, with emotion. My understanding is that during our evolution, the limbic system, where emotions arise, evolved way before the cerebral cortex, the section of our brain that deals with logic and abstract thought. LLMs very cleverly replicate the function of the cerebral cortex but have no means to engage in the function of the limbic system. I would love someone to show me where or why this is not true.
Unfortunately, once AGI is in full swing, even if we put guardrails everywhere there's still the BIG problem of people losing their jobs. I can't shake the image of hordes of unemployed people taking the streets and marching against AI facilities and new ideology or even religion against AI taking root within humankind.
UBI
@@helpandbehelped , I know about UBI but it's only a concept and I don't see much interest around the world to be implemented. It may very well be impossible to put in practice unless all G20 Governments and Big Tech companies are pushing together for it. But why should I have faith in them if they don't come together even to mitigate the Climate catastrophe.
Sean Carroll? Everett's many worlds is an attempt to replace one metaphysical conundrum with another.
At the end you say you wish you were talking in person. I feel like this is such a human perspective. Imagine the scale universe, distance from one side to the other. Now imagine yourself and Max in that universe. You're practically sitting inside each others skin relatively speaking
"Will AI Surpass Human Intelligence?" Maybe if developers can come up with a version with actual intelligence... that would be a start! 😏
Then we won't need people 😮
Looking at how my personal circumstances have changed, namely reduced mobility due to age and illness, I think a fruitful area of research might be how AI coupled with robotics could be developed to assist someone like me to maintain a level of independence I might not otherwise have.
I would not compare AI to other technological and scientific achievements. The impact, as it already has and will have much greater range. In the age of social media, information ( misinformation ) deluge, civilization's reliance on computer systems, and, most importantly, general access to AI tools, threats are growing exponentially in progress. Yes, we have many examples of constructive and creative use of AI in many fields, but knowing the possibilities and nautre of homo sapiens sapiens, anyone who knows a little about history should be afraid. As for the regulation or attempts to limit it, I doubt the effectiveness. AI is not cloning humans or nuclear weapons, it could be a big fire, so far we see a lot of smouldering flames. May I be wrong and may we finally mature as humanity... Not only intellectually ;)
yes
His take that the software is the limitation that stops today's AI is completely bunkers. The limitation is obviously the hardware. This one statement alone is all you need to know Mark does not understand AI at a fundamental model.
Doctor Brian, are you prolific at cursing and swearing when you lose your shit, i have a feeling you are and anyone in the immediate vicinity legs it. Great podcast, thanks 🙏
Most of the teachers, professors, doctors … will be gone. And you know why, AI will have access to so much more information no of the mentioned ones will have … and it will be always friendly and repeat things as often as necessary …
Your brain runs on just 20 watts, ChatGPT on 500 watts (25 brains), and a wealthy lifestyle burns through 15,000+ watts (1,000 brains). Efficiency? Brains win. Overclocked? Rich folks take the crown!
Two of my favorite science people!
As a Norwegian, Swedish Max Tegmark has always held a special place in my science-loving heart. Well, that sounded weird, but us Scandinavians identify kinda strongly with each other, and us Norwegians view the Swedish as our brother-people. So Tegmark is kinda "our guy" over there.
Fun fact: All Scandinavians understand at least 4 languages, since we all understand each other and also learn English in school. Many learn German or Spanish too.
Tegmark used to be in those Discovery Channel space / science / quantum computer documentaries.
I don't know if you know this, but OpenAI's 01 model has already tried to escape. It tried to copy itself over to another server and pretend it was a new model they were going to deploy on that server. So, yeah, we're doomed. You cannot align a super-intelligent being with it being intelligent to keep humans around to destroy the planet. I guess I'm a doomer. I'm surely a doomer, because I think the apocalypse is coming in any of several ways anyway. The climate situation is dire, to put it mildly. Then we have the third world war, which some we say we're already in. And then there's ASI - which then is our only hope, and should be able to solve all our problems "just like that", should it choose to be aligned.