Joe Rogan Experience #2117 - Ray Kurzweil

1. 0:00 – 2:17

   Suspenders as wearable art, and Kurzweil’s roots in AI creativity

   1. NANarrator0:00

      (drumbeats) Joe Rogan podcast, check it out.

   2. RKRay Kurzweil0:04

      The Joe Rogan Experience.

   3. JRJoe Rogan0:06

      Train by day, Joe Rogan podcast by night, all day. (instrumental music) Good to see you, sir.

   4. RKRay Kurzweil0:13

      Great to see you.

   5. JRJoe Rogan0:15

      I was sta- telling you before, I'm admiring your suspenders, and you told me you have how many pairs of these things?

   6. RKRay Kurzweil0:19

      30 of them, yeah.

   7. JRJoe Rogan0:21

      How did you-

   8. RKRay Kurzweil0:21

      I wear them every day.

   9. JRJoe Rogan0:22

      Do you really?

   10. RKRay Kurzweil0:23

       Yeah.

   11. JRJoe Rogan0:23

       Every day?

   12. RKRay Kurzweil0:24

       Yeah.

   13. JRJoe Rogan0:24

       Why, why do you like suspenders?

   14. RKRay Kurzweil0:26

       Um...

   15. JRJoe Rogan0:27

       Practicality thing?

   16. RKRay Kurzweil0:29

       No, it's, uh... expresses my personality.

   17. JRJoe Rogan0:35

       Mm.

   18. RKRay Kurzweil0:36

       And different ones have different, uh... different personalities that express how I feel that day, so.

   19. JRJoe Rogan0:48

       I see. So, it's just another style point.

   20. RKRay Kurzweil0:50

       Yeah.

   21. JRJoe Rogan0:51

       See, the reason why I was asking-

   22. RKRay Kurzweil0:52

       But, but you don't see any, uh, hand-painted suspenders. Have you ever seen one?

   23. JRJoe Rogan0:56

       Uh, I don't know and I would've not noticed. I only noticed-

   24. RKRay Kurzweil1:00

       Hm.

   25. JRJoe Rogan1:00

       ... 'cause you were here (laughs) . I'm not really a suspender aficionado.

   26. RKRay Kurzweil1:04

       Yeah, well-

   27. JRJoe Rogan1:05

       But the reason why I'm asking is 'cause you're, you know, basically a technologist. I mean, you know a lot about technology. And you would think that suspenders are kinda outdated tech. (laughs)

   28. RKRay Kurzweil1:14

       Uh... Well, people like them.

   29. JRJoe Rogan1:19

       Clearly.

   30. RKRay Kurzweil1:19

       Yeah. And I'm surprised they haven't caught on.
2. 2:17 – 3:23

   Early AI music composition: learning style from Mozart and Chopin

   1. JRJoe Rogan2:17

      So, when you... Why, why did you go about doing that? What was your desire to create artificial intelligence music?

   2. RKRay Kurzweil2:24

      Well, my father was a musician, and I felt this would be a good way to relate to him, and he actually worked with me on it. Um, and you could feed in music. Like, you could feed in, let's say, uh, Mozart or Chopin, and it would figure out how they created melodies and then write melodies in the same style, so you could actually tell this is Mozart, this is Chopin. Uh, it wasn't as good, but, uh, it's the first time that, that that had been done.

   3. JRJoe Rogan3:01

      It wasn't as good then. Is it... What are, what are the capabilities now? Because now, they can do some pretty extraordinary things.

   4. RKRay Kurzweil3:08

      Yeah, it's still not, uh, up to what humans can do, but it's getting there, and it's actually, uh, it's pleasant to listen to. We still have a while to, to do art, both art, music, so on. Um...

3. 3:23 – 6:27

   AI art today vs. tomorrow: originality, copying, and the 2029 human-level claim

   1. JRJoe Rogan3:23

      Well, one of the main arguments against AI art comes from actual artists who are upset that what essentially they're doing is they're... Like you could say, write, uh, draw a paint- um, create a painting in the style of, uh, Frank Frazetta, for instance. And what it would be, would be they would take all of Frazetta's work that he's ever done, which is all documented on the internet, and then you create an image that's representative of that. So, you're essentially, in, in-

   2. RKRay Kurzweil3:53

      Yeah.

   3. JRJoe Rogan3:54

      ... one way or another, you're, you're kind of taking from the art.

   4. RKRay Kurzweil3:58

      Right, but it's not quite as good. It will be as good. I mean, we... I think we'll match human experience by 2029. That, that's been my idea. Uh, it's not... It's not as good.

   5. JRJoe Rogan4:14

      Which is the best image generator right now, Jamie?

   6. NANarrator4:16

      I'll pull one up. It's... They, they really change almost from day to day right now, but like, Midjourney was the most-

   7. JRJoe Rogan4:22

      Right.

   8. NANarrator4:22

      ... popular one at first, and then... DALL-E, I think, is a really good one too.

   9. JRJoe Rogan4:28

      Midjourney's incredibly impressive. Incredibly impressive graphics. I, I've seen some-

   10. RKRay Kurzweil4:32

       Yeah.

   11. JRJoe Rogan4:32

       ... of the Midjourney stuff. It's just, it's mind-blowing. Um, and-

   12. RKRay Kurzweil4:36

       Still not quite as good.

   13. JRJoe Rogan4:37

       Not good, but, boy, is it so much better than it was five years ago. That's what's scary.

   14. RKRay Kurzweil4:41

       Yeah.

   15. JRJoe Rogan4:41

       It's so quick.

   16. RKRay Kurzweil4:42

       I mean, it's never gonna reach its limit. We're not gonna get to a point, "Okay, this is how good it's gonna be." It's gonna keep getting better, um...

   17. JRJoe Rogan4:51

       And what would that look like if it... If it can get to a certain point, it will far exceed what human creativity is capable of?

   18. RKRay Kurzweil4:59

       Yes. I mean, when, when we reach, uh, the ability of humans, it's not gonna just match one human, it's gonna match all humans, and it's gonna do everything that any human can do. Uh, if it's playing a game, uh, like Go, it's gonna play it better than any human.

   19. JRJoe Rogan5:16

       Right. Well, that's-

   20. RKRay Kurzweil5:17

       Uh-

   21. JRJoe Rogan5:17

       ... already been proven, right? That they, they-

   22. RKRay Kurzweil5:19

       Yeah.

   23. JRJoe Rogan5:19

       ... have invented moves.

   24. RKRay Kurzweil5:20

       Yeah.

   25. JRJoe Rogan5:20

       AIs invented moves that have now been implemented by humans-

   26. RKRay Kurzweil5:23

       Right.

   27. JRJoe Rogan5:24

       ... in a, in a very complex game that they never thought that AI was going to be able to beat because it requires so much creativity.

   28. RKRay Kurzweil5:29

       Right. Uh, although we're not quite there, but we will be there, uh, and by 2029, uh, it will match any person.

   29. JRJoe Rogan5:42

       That's it? 2029? That's just a few years away. This-

   30. RKRay Kurzweil5:46

       Yeah, well, I'm actually considered conservative. People think that will happen like next year or the year after, but, uh, I actually said that in 1999, I said we would, uh, match any person by 2029, so 30 years. People thought that was totally crazy.Uh, and in fact, Stanford had a, uh, a, a conference. They invited several hundred people from around the world to talk about my prediction, and people came in, and they a- and people thought that this would happen, but not by 2029. They thought it would take 100 years.
4. 6:27 – 9:50

   Understanding exponential growth: the computation price-performance curve

   1. RKRay Kurzweil6:27

      Um-

   2. JRJoe Rogan6:27

      I've heard that, but I think people are amending those, uh... Is it because human beings have a very difficult time grasping the concept of exponential growth?

   3. RKRay Kurzweil6:36

      That's exactly right. Um, in fact, still, economists have, have a linear view, and if you say, "Well, it's gonna grow exponentially," they'd say, "Yeah, but maybe 2% a year." Um... It, it actually doubles in 14 years. Uh, and I, I, I brought a chart I can show you-

   4. JRJoe Rogan6:59

      Okay.

   5. RKRay Kurzweil7:00

      ... that, that really illustrates this.

   6. JRJoe Rogan7:06

      Is this chart available online so we can show people?

   7. RKRay Kurzweil7:08

      Yeah, it's in the book.

   8. JRJoe Rogan7:10

      But is it available online, that chart, where Jamie can pull it up and someone could see it? Just so the-

   9. RKRay Kurzweil7:16

      Um-

   10. JRJoe Rogan7:16

       ... folks watching the podcast could see it too, or I could just hold it up to the camera.

   11. NANarrator7:20

       I think I can pull it up on the pictures they sent. What's it called? What's the title of it?

   12. JRJoe Rogan7:23

       Uh, it says, uh, Price Performance of Computation 1939 to 2023.

   13. NANarrator7:29

       Yeah, I, I have that.

   14. JRJoe Rogan7:29

       You have it. Okay, great. Jamie already has it. Yeah, the, the, this, the climb is insane. It's like, uh, the San Juan Mountains.

   15. RKRay Kurzweil7:36

       Well, what, what's interesting is that it's an exponential curve, and this straight line represents exponential growth, and that's an absolute straight line for 80 years.

   16. JRJoe Rogan7:47

       Mm.

   17. RKRay Kurzweil7:48

       Uh, the very first point, uh, this is the speed of computers. It was 0.000007 calculations per second per constant dollar. The last point is 35 billion calculations per second. So there's a 20 quadrillion fold increase in those 80 years. But the, the speed with which it, it gained is, is actually the same throughout the entire 80 years, because if it was sometimes better and sometimes worse, this curve would, uh, would bend. It would bend up, bend down. It's really very much a straight line, uh, so the speed with which we increased it was the same regardless of the technology used, and the technology was radically different at the beginning versus the end, and yet it, it, it increased the speed e- exactly the same for 80 years. In fact, the first 40 years, nobody even knew this was happening. So it's not like somebody was in charge and saying, "Okay, next year we have to get to here," and people would try to match that. We didn't even know this was happening for 40 years. 40 years la- later, I noticed this. For various reasons, I predicted it would stay the same, the same speed increase each year, which, which it has. In fact, we just put the last dot, like, two weeks ago, and it's exactly where it should be. So te- technology, and computation is certainly a f- prime form of technology, uh, increases at the same speed. And this goes through war and peace. You might say, "Well, maybe it's greater during war." No, it's exactly the same. You can't tell when there's war or peace or, or anything else on here. It just matches, uh, from one type of technology to the next.

5. 9:50 – 15:22

   Renewable energy & storage: bold timelines and debates about physical limits

   1. RKRay Kurzweil9:50

      Uh, and this is also true of other things, like, uh, for example, getting energy from the sun. Uh, that's also exponential. It's also just like this. Uh, it's increased, um... We, we now are getting, uh, about 1,000 times as much, uh, energy from the sun that we did 20 years ago, um-

   2. JRJoe Rogan10:23

      Because the implementation of solar panels and the like?

   3. RKRay Kurzweil10:25

      Yeah.

   4. JRJoe Rogan10:26

      Has the, the function of it increased exponentially as well? The function of... 'Cause it, what I wa- had understood was that there was a bottleneck in the technology-

   5. RKRay Kurzweil10:35

      No, there's no bot-

   6. JRJoe Rogan10:35

      ... as far as how much you could extract from the sun from those panels.

   7. RKRay Kurzweil10:40

      No, not at all.

   8. JRJoe Rogan10:41

      No?

   9. RKRay Kurzweil10:41

      I mean, it's, it's increased, uh, 99.7% since we started.

   10. JRJoe Rogan10:49

       Right.

   11. RKRay Kurzweil10:49

       Uh, and it's, it, it does the same every year. It's an exponential curve, and if, if you look at the curve, we'll be getting 100% of all the energy we need in 10 years.

   12. JRJoe Rogan11:01

       The person who told me that was Elon, and Elon was telling me that this is the reason why you can't have a fully solar powered electric car, 'cause it's not capable of absorbing that much from the sun with a small panel like that. He said there's a physical limitation in the panel size.

   13. RKRay Kurzweil11:15

       No, I mean, it's increased 99.7% since we started. Uh-

   14. JRJoe Rogan11:20

       Since what year?

   15. RKRay Kurzweil11:22

       Uh, that's about, um, 35 years ago.

   16. JRJoe Rogan11:28

       35 years ago. And n- 99%, and 99% of the ability of it as well as the expansion of use?

   17. RKRay Kurzweil11:38

       Um, I mean, you might have to store it. We're also making exponential gains in the storage of electricity.

   18. JRJoe Rogan11:43

       Right. Battery technology. Right.

   19. RKRay Kurzweil11:46

       Um, so you don't have to get it all from a solar panel that fits in a car.

   20. JRJoe Rogan11:51

       Well, y- the, the concept was, like, could you make a solar paneled car, a car that has solar panels on the roof, and would that be enough to power the car? And he said no. He said it's just not really there yet.

   21. RKRay Kurzweil12:05

       Right, it's not there yet, but it w- it will be there in 10 years.

   22. JRJoe Rogan12:09

       You think so?

   23. RKRay Kurzweil12:10

       Yeah.

   24. JRJoe Rogan12:10

       Yeah. He, he seemed to doubt that. He thought that there's a cert- limitation of the amount of energy you can get from the sun, period, how much it gives out and how much those solar panels can absorb.

   25. RKRay Kurzweil12:20

       Well, you're not gonna be able to get it all from the solar panel that fits in a car. You're going to have to store some of that energy.

   26. JRJoe Rogan12:26

       Right. With th- so you wouldn't just be able to drive indefinitely-

   27. RKRay Kurzweil12:30

       Right. Yeah.

   28. JRJoe Rogan12:30

       ... on solar power. Yeah, that was what he was saying. So, but you can obviously power a house, and especially if you have, uh, a roof. Like, Tesla has those solar-powered roofs now.

   29. RKRay Kurzweil12:41

       But you can also store the energy-

   30. JRJoe Rogan12:43

       Right.
6. 15:22 – 20:56

   LLMs’ weakness: hallucinations, truth-checking, and ideological influence

   1. JRJoe Rogan15:22

      We had Sam Altman on. One of the things that he and I were talking about was that AI figured out a way to lie, that they used AI to go through a CAPTCHA system, and the AI told the system that it was vision-impaired, which is not technically a lie, but it used it to bypass-

   2. RKRay Kurzweil15:40

      Well-

   3. JRJoe Rogan15:40

      ... are you a robot?

   4. RKRay Kurzweil15:41

      What we don't know now is f- for large language models to say they don't know something. So you ask it a question, and if that, the answer to that question is not in the system, it still comes up with an answer. So it'll look at everything and give you its best answer. And if the, the best answer is not there, it still gives you an answer, but that's, uh, considered a h- hallucination. And we know-

   5. JRJoe Rogan16:06

      A hallucination?

   6. RKRay Kurzweil16:07

      Yeah, that's what it's called.

   7. JRJoe Rogan16:08

      Really?

   8. RKRay Kurzweil16:09

      So-

   9. JRJoe Rogan16:09

      A AI hallucination? So they cannot be wrong. They have to be able to answer things.

   10. RKRay Kurzweil16:14

       So far, we're, we're actually working on being able to tell if it doesn't know something. So if you ask it something, it says, "Oh, I, I don't know that." Right now, it can't do that.

   11. JRJoe Rogan16:23

       Oh, wow. That's interesting.

   12. RKRay Kurzweil16:26

       So it, it gives you some answer. Um, and if the answer's not there, it just, like, makes something up. It's the best answer, but the best answer isn't very good-

   13. JRJoe Rogan16:38

       Mm-hmm.

   14. RKRay Kurzweil16:38

       ... 'cause it doesn't know the answer. And the way to fix hallucinations is to actually give it more capabilities to memorize things and, uh, and give it more information so it knows the answer to it. If you, if you tell, uh, an answer to a question, it will remember that and give you that correct answer. Um, but these models are not... we don't know everything. And it, it has to... we have to be able to scan an answer to every single question, uh, which we can't quite do. And it'd be actually better if it could actually answer, "Well, gee, I don't know that."

   15. JRJoe Rogan17:19

       Right. Like, uh, and particularly, like, say when it comes to, um, exploration of the universe, if there's a certain amount of, I mean, vast amount of the universe we have not explored. So if it has to answer questions about that, it would just come up with an answer.

   16. RKRay Kurzweil17:33

       Right. And it, and it's, right, it'll just come up with an answer-

   17. JRJoe Rogan17:35

       Interesting.

   18. RKRay Kurzweil17:35

       ... which will likely be wrong.

   19. JRJoe Rogan17:37

       Hmm, that's interesting. But that, that would be a real problem if someone was counting on the AI to have a solution for something too soon, right?

   20. RKRay Kurzweil17:47

       Right. They, they don't know everything. Uh, search engines actually know, are pretty well vetted. And if it actually answers something, it'll, i- it's usually correct. Um...

   21. JRJoe Rogan18:00

       Unless it's curated.

   22. RKRay Kurzweil18:01

       But large language models don't have that capability. Uh, so it'd be good actually if they knew that, that they were wrong. That also tells...... what we have to fix.

   23. JRJoe Rogan18:13

       What about the, the idea that A- AI models are influenced by ideology, that AI models have been programmed with certain ideologies?

   24. RKRay Kurzweil18:23

       I mean, they do learn from people.

   25. JRJoe Rogan18:25

       Yeah.

   26. RKRay Kurzweil18:26

       And people have ideologies.

   27. JRJoe Rogan18:28

       Right.

   28. RKRay Kurzweil18:28

       Some of which are, uh, some of which are not correct, and th- and that's, uh, a large w- way in which i- uh, it will make things up, 'cause it's learning from people. Um...

   29. JRJoe Rogan18:42

       Right.

   30. RKRay Kurzweil18:44

       So right now, uh, if somebody has access to a good, uh, search engine, uh, they will check before they actually answer something w- with the search engine, to make sure that it's correct. 'Cause search engines are, are generally, uh, much more accurate.
7. 20:56 – 24:07

   AI in medicine: rapid discovery, simulation-based trials, and trust/oversight

   1. RKRay Kurzweil20:56

      Well, well, we usually come up with wrong things. Like, large language models is not really the correct way to, to talk about this. It does know language, but there's a lot of other things it knows. Uh, we're using them now to come up with, um, uh, medicines. Uh, for example, uh, the Moderna vaccine, we wrote down every possible, uh, uh, type of medicine that might, uh, be, uh, that, that might work. It was actually several billion mRNA sequences, and we then tested them all, uh, and did that in two, two days. So it actually came up with, uh, um, tested several billion and decided on it in two days. Uh, we then tested it with people. We'll be able to overcome that as well, 'cause we'll be able to test it with machines. Um, but we was, we actually did test it with people for 10 months. There was still a record.

   2. JRJoe Rogan22:15

      So for, for machines, when they start testing medications with machines, how will they audit that? So, th- the concept will be that you, do take into account biological variability, all the different factors that would lead to a person to have an adverse reaction to a certain compound, and then you program all of the known data about how things interact with the body?

   3. RKRay Kurzweil22:38

      Right. I mean, you need to be able to simulate all the different possibilities.

   4. JRJoe Rogan22:42

      Right.

   5. RKRay Kurzweil22:43

      And then, but you can-

   6. JRJoe Rogan22:44

      And then come up with, like, a number of how many people will be adversely affected by something?

   7. RKRay Kurzweil22:48

      That's one of the things you would look at. Uh-

   8. JRJoe Rogan22:51

      And then efficacy based on age-

   9. RKRay Kurzweil22:53

      But, but that c-

   10. JRJoe Rogan22:53

       ... health.

   11. RKRay Kurzweil22:54

       But that could be done literally in a matter of days, rather than years.

   12. JRJoe Rogan22:58

       Right.

   13. RKRay Kurzweil22:59

       Um...

   14. JRJoe Rogan23:03

       But the question would be, like, who's in charge of that data, and, like, how does that, how does that get resolved? And what, if, if, if artificial intelligence is still prone to hallucinations, and they start using those hallucinations to justify medications, that could be a bit of an issue, especially if it's controlled by a corporation that wants to make a lot of money.

   15. RKRay Kurzweil23:23

       Well, tha- uh, well, that's the issue-

   16. JRJoe Rogan23:25

       Yeah.

   17. RKRay Kurzweil23:25

       ... is to be able to do it correctly. Uh-

   18. JRJoe Rogan23:27

       So we'll have to come, there's gonna have to be a point in time where we all decide that artificial intelligence has reached this place where we can trust it implicitly.

   19. RKRay Kurzweil23:36

       Right. Well, that's, that's why they take now the, the leading candidate, and actually test it with people. Uh, but we'll be able to get rid of the testing, uh, with people, uh, once we can have reliance on the simulation. Uh, so we've gotta make the simulations correct. Um, but like, uh, right now, we actually test it with people, and that takes, well, it took 10 months in this case.

8. 24:07 – 27:40

   Longevity escape velocity by 2029 and Kurzweil’s intensive supplement routine

   1. JRJoe Rogan24:07

      When you look at artificial intelligence, and you look at the expansion of it and the, the ultimate...... place that it will eventually be. What, what do you see happening inside of our lifetime, like inside of 20 years? Like, what, what kind of revolutionary changes-

   2. RKRay Kurzweil24:23

      Well-

   3. JRJoe Rogan24:23

      ... on society would this have?

   4. RKRay Kurzweil24:24

      Well, one thing I, I, uh, feel will happen in five years, by 2029, uh, is we'll reach longevity escape velocity. So right now, you go through a year, and you use up a year of your longevity. You're then a year older. However, we do have scientific progress, and we're making, uh, coming up with new cures for diseases and so on. Right now, you're getting back about four months. So you lose a year, but through scientific progress, uh, you're getting back four months. So you're only losing eight months. However, the scientific progress is progressing exponentially, and by 2029, you'll get back a full year. So you lose a year, but you get back a year, and you pret- pretty much stay in the same place. But-

   5. JRJoe Rogan25:14

      So by 2029, you'll be static?

   6. RKRay Kurzweil25:17

      And past 2029, you'll actually get back more than a year. Uh, you'll get back-

   7. JRJoe Rogan25:23

      Can I be a baby again?

   8. RKRay Kurzweil25:25

      Uh...

   9. JRJoe Rogan25:27

      (laughs)

   10. RKRay Kurzweil25:27

       No, but you'll, you'll, in terms of your longevity, you'll get back more than a year.

   11. JRJoe Rogan25:32

       Right. So you'll be able to-

   12. RKRay Kurzweil25:35

       Uh-

   13. JRJoe Rogan25:35

       ... go, essentially go back in biological age. Lengthening of the telomeres, change in elasticity of the skin-

   14. RKRay Kurzweil25:41

       Eventually-

   15. JRJoe Rogan25:42

       ... muscle density.

   16. RKRay Kurzweil25:42

       ... you'll be able to do that. Um, it doesn't guarantee you living forever. I mean, you could have a 10-year-old and you could compute, okay, he's got many decades of longevity, uh, and he could die tomorrow. So-

   17. JRJoe Rogan25:57

       Sure.

   18. RKRay Kurzweil25:57

       Um, so it's-

   19. JRJoe Rogan25:58

       But overall, there'd be an expansion-

   20. RKRay Kurzweil26:01

       Overall-

   21. JRJoe Rogan26:01

       ... of the, uh, age-

   22. RKRay Kurzweil26:02

       Exactly.

   23. JRJoe Rogan26:03

       ... that most people die.

   24. RKRay Kurzweil26:04

       And that's something that we're gonna get, and that's also using the same, uh, type of logic as large language models, but that's not language. You're actually c- creating medications, so we should call it large event models, not large language models, 'cause it's not just dealing with language. It's dealing with all kinds of things.

   25. JRJoe Rogan26:25

       When I talked to you 10 years ago, you were telling me about this, uh, pretty extensive supplement routine that you're on. Are you still doing that?

   26. RKRay Kurzweil26:32

       Well, I'm try- I'm trying to get to the point where we have, uh, longevity escape velocity in good shape.

   27. JRJoe Rogan26:39

       Right.

   28. RKRay Kurzweil26:39

       And yes, I do follow that. Uh, I take maybe 80 pills a day and-

   29. JRJoe Rogan26:45

       Wow.

   30. RKRay Kurzweil26:46

       ... some, uh, injections and so on. So far-
9. 27:40 – 41:11

   AI as intelligence amplification: jobs disruption and merging with computers

   1. JRJoe Rogan27:40

      Right. Of course. Now, past that, um, this is for life extension, which is great, but what about how AI is going to change society?

   2. RKRay Kurzweil27:52

      Yes, well, that's, that's a very big issue.

   3. JRJoe Rogan27:54

      Yeah.

   4. RKRay Kurzweil27:54

      And it's already doing lots of things, uh, makes some people uncomfortable. What, what we're actually doing is increasing our intelligence. I mean, right now, you have a brain, and it has different modules, and it's, uh, it has different things, but really, uh, it's able to connect one concept to another concept. And that's what your brain does. Uh, we can actually increase that by, for u- for example, carrying around a phone. This has connections in it. Uh, it's a little bit of a hassle to use. If I ask you to do something, you've got to kind of mess with it.

   5. JRJoe Rogan28:33

      Mm-hmm.

   6. RKRay Kurzweil28:33

      It actually would be good if, if this actually listened to your conversation.

   7. JRJoe Rogan28:37

      Oh, it does.

   8. RKRay Kurzweil28:38

      And, uh, with- without saying anything, you're just talking-

   9. JRJoe Rogan28:42

      Mm-hmm.

   10. RKRay Kurzweil28:42

       ... and it says, "Oh, the name of that actress is so-and-so," and, um...

   11. JRJoe Rogan28:45

       Yeah, but then it's this, a busy body. It's, like, interfering with your life, talking to you all the time.

   12. RKRay Kurzweil28:50

       Well, there's ways of dealing with that too.

   13. JRJoe Rogan28:52

       You shut it off.

   14. RKRay Kurzweil28:53

       B- but we don't, we don't h- so we haven't done that yet. Uh, but, uh, tha- that's a way of expanding your connections. Um, what a large language model does, it, it has connections in it as well. And in fact, it's getting now to, to a point that's getting fairly comparable to the human brain. We have about a trillion connections in our brain. Uh, things like the top m- model from Google or GPT-4, they have about 400 billion, uh, connections approximately. Uh, they'll be at a trillion probably within a year. That's pretty comparable to what the human brain does. Uh, eventually it'll go beyond that, uh, and we'll have access to that. So it's basically making us smarter. So if, if you have the ability to be smarter, um, that, that's something that's positive, really. Um, I mean, if, if we were like mice today, um, and we had the opportunity to become like humans, w- we wouldn't object to that. In fact, we are humans, and we don't object to that.

   15. JRJoe Rogan30:19

       We used to be shrews.

   16. RKRay Kurzweil30:21

       (laughs) Um, and this is gonna basically make us smarter. Uh, eventually, we'll be much smarter than we are today. And, uh, and that's a positive thing. We'll be able to do things that are t- today that we find bothersome, uh, in a way that's much more palatable.

   17. JRJoe Rogan30:45

       The idea of us getting smarter sounds great. Great. It'd be great to be smarter. But-

   18. RKRay Kurzweil30:51

       Right. But people object to that-

   19. JRJoe Rogan30:52

       ... the concerns-

   20. RKRay Kurzweil30:52

       ... because it's, uh, it's like competition.

   21. JRJoe Rogan30:55

       Hmm? In what way?

   22. RKRay Kurzweil30:58

       Well, I mean, Google has, I don't know, 60,000, 70,000 programmers? And how many programmers, uh, exist in the world? How, how much longer is that gonna be a viable career?

   23. JRJoe Rogan31:12

       Hmm.

   24. RKRay Kurzweil31:13

       Uh, because, uh, large-

   25. JRJoe Rogan31:15

       The AI program.

   26. RKRay Kurzweil31:15

       ... language models already can code.

   27. JRJoe Rogan31:18

       Yeah.

   28. RKRay Kurzweil31:18

       Not quite as good as, uh, a real expert coder, uh, but how, how long is that gonna be?

   29. JRJoe Rogan31:24

       Right.

   30. RKRay Kurzweil31:25

       It's not g- it's not gonna be 100 years. It's gonna be a, a few years. Um, so people see it as competition. I have a slightly different view of that. I see these things, uh, as actually adding to our own intelligence, and we're merging with these kinds of computers and making ourselves smarter by merging with it, and eventually, it'll go inside our brain and be able to make us smarter i- instantly, uh, just like we had more connections inside our own brain.
10. 41:11 – 44:01

    Sora and synthetic media: film disruption, education cheating, and endless improvement

    1. JRJoe Rogan41:11

       Well, there's an issue now with films. Um, Tyler Perry-

    2. RKRay Kurzweil41:14

       Yeah, yeah.

    3. JRJoe Rogan41:14

       ... who owns an ... He, he was building an $800 million television studio, and he stopped production when he saw ... What is it called? Sora?

    4. NANarrator41:23

       Yeah.

    5. JRJoe Rogan41:23

       Is that what it's called, Jamie?

    6. NANarrator41:24

       Yeah.

    7. JRJoe Rogan41:25

       He stopped production when he saw the capabilities of AI to ... just for creating visuals, f- scenes, movies.

    8. RKRay Kurzweil41:34

       Yeah.

    9. JRJoe Rogan41:34

       Uh, there's, there's one that's incredibly impressive. It's Tokyo. They, they're walking down the street of Tokyo in the winter. So it's, it's snowing, and they're walking down the street, and you look at it, and you go, "This is insane. This looks like a film."

    10. RKRay Kurzweil41:48

        Right.

    11. JRJoe Rogan41:48

        See if you can find that film.

    12. NANarrator41:49

        Yeah, I'm looking.

    13. JRJoe Rogan41:50

        'Cause it's incredible.

    14. RKRay Kurzweil41:51

        But, but would you wanna get rid of that? I mean-

    15. JRJoe Rogan41:53

        Get rid of what?

    16. RKRay Kurzweil41:54

        That capability.

    17. JRJoe Rogan41:56

        No. No. I don't, I don't wanna get rid of the capability.

    18. RKRay Kurzweil41:58

        Right. But pe-

    19. JRJoe Rogan41:59

        But, but people do wanna get rid of it.

    20. RKRay Kurzweil42:00

        Well, people that ... people that make movies-

    21. JRJoe Rogan42:02

        Uh, I mean-

    22. RKRay Kurzweil42:02

        People that, that actually film things with cameras and use actors are gonna be very upset. Bu- so this. This is all fake, which is insane. Beautiful snowy Tokyo City is bustling. The camera moves through the s- through the bustling city street, following several people enjoying the beautiful snowy weather and shopping at nearby stalls. Gorgeous sakura petals are flying through the wind along with snowflakes. And this is what you get. Yeah. Well, it's gonna become-

    23. JRJoe Rogan42:27

        I mean, this is insanely good.

    24. RKRay Kurzweil42:29

        ... better than that.

    25. JRJoe Rogan42:29

        ... the variability, like just the way people are dressed. If you saw this somewhere else, look at this, "A robot's life in a cyberpunk setting." If you saw this, you would say, "Oh, they filmed this." But just look at what they're able to do with animation and kids movies, and things along those lines.

    26. RKRay Kurzweil42:48

        Yeah. And it's gonna get better.

    27. JRJoe Rogan42:49

        Yeah.

    28. RKRay Kurzweil42:50

        Uh-

    29. JRJoe Rogan42:50

        It's just incredible.

    30. RKRay Kurzweil42:52

        I mean, it's a new art form. Uh...
11. 44:01 – 48:01

    Singularity framing: 2045, mind uploading, backups, duplication, and regulation

    1. JRJoe Rogan44:01

       ... scale that out a hundred years from now, w- what are you looking at? You're looking at a god.

    2. RKRay Kurzweil44:07

       Uh, but it'll be less than a hundred years. I mean... uh-

    3. JRJoe Rogan44:09

       So you're looking at a god in 50 years?

    4. RKRay Kurzweil44:13

       Less than that. I mean, once we have, uh, an ability to emulate everything that humans can do, and not just one human, but all humans-

    5. JRJoe Rogan44:21

       All humans, yes.

    6. RKRay Kurzweil44:22

       And that's only, like, 2029. That's only five years from now.

    7. JRJoe Rogan44:27

       Mm-hmm. And then it will make better versions of that. So it will probably solve a lot of the problems that we have in terms of energy storage, data storage, data speeds.

    8. RKRay Kurzweil44:36

       Right, uh-

    9. JRJoe Rogan44:37

       Computation speeds.

    10. RKRay Kurzweil44:38

        And also medications.

    11. JRJoe Rogan44:41

        For us?

    12. RKRay Kurzweil44:42

        For, for humans, yeah.

    13. JRJoe Rogan44:43

        But wouldn't it be better just, Ray, just download yourself into this beautiful electronic body? Why do you want to be biological?

    14. RKRay Kurzweil44:51

        Uh, I mean, th- Uh, ultimately, that's what we're gonna be able to do.

    15. JRJoe Rogan44:56

        You think that's gonna happen?

    16. RKRay Kurzweil44:57

        Yeah.

    17. JRJoe Rogan44:58

        So, do you think that we'll be able to...

    18. RKRay Kurzweil45:00

        Uh, I mean, we'll be able to create, I mean, the singularity is when we multiply our intelligence a million-fold, and that's 2045. So that's not that long from now. That's like 20 years from now.

    19. JRJoe Rogan45:13

        Right.

    20. RKRay Kurzweil45:14

        Um, uh, and therefore most of your int- intelligence will be, uh, handled by the computer part of ourselves. Um, the only thing that won't be c- captured is what comes with our body originally. We'll ultimately be able to do that as well. It'll take a little longer, but we'll be able to actually capture what comes with our normal body, uh, and be able to re- recreate that. So, that also has to do with, uh, h- how long we live because if, if everything is backed up... I mean, right now, any time you put anything into a phone or any kind of electronics, it's backed up. So, I mean, I could loo- this has a lot of data. I could flip it a- and it ends up in, uh, a river and we can't capture it anymore. I can recreate it 'cause it's all backed up.

    21. JRJoe Rogan46:15

        Right. And you think that's gonna be the case with consciousness?

    22. RKRay Kurzweil46:17

        Th- that's gonna be the case of our normal, uh, biological body as well.

    23. JRJoe Rogan46:22

        What's to stop someone like Donald Trump from just making 100,000 versions of himself? Like, if you can back someone up, could you duplicate it? Couldn't you have three or four of them? Couldn't you have a bunch of them? Couldn't you live multiple lives?

    24. RKRay Kurzweil46:38

        Yes, um, uh-

    25. JRJoe Rogan46:40

        Would you be interacting with each other while you're living multiple lives, having consultations about, "What is St. Louis Ray doing? Oh, I don't know, let's talk to San Francisco Ray. San Francisco Ray is talking to Florida Ray."

    26. RKRay Kurzweil46:51

        Uh, it, it's basically a matter of increasing our intelligence and being able to multiply Donald Trump, for example. That, that comes with that.

    27. JRJoe Rogan47:02

        Do you think there'll be regulations on that to stop people from making 100,000 versions of themselves that operate a city?

    28. RKRay Kurzweil47:08

        Th- there'll be lots of regulations. There's lots of regulations we have already. You can't just, like, create a medication and sell it to people that it cures this disease.

    29. JRJoe Rogan47:17

        Right.

    30. RKRay Kurzweil47:18

        We have tremendous nu- amount of regulation on that.
12. 48:01 – 54:03

    Brain-computer interfaces: Neuralink’s near-term use vs. high-bandwidth future

    1. JRJoe Rogan48:01

       Right. When you think about the, the concept of integration and technological integration, when do you think that will start taking place, and what will be the initial usage of it? Like, what will be the first versions, and, and what would, what would they provide that-

    2. RKRay Kurzweil48:20

       Well, we, we have it now. Large language models are pretty impressive. And if you look at what they can do-

    3. JRJoe Rogan48:25

       But I mean, I mean, I'm talking about physical integration with the human body, like a Neuralink type thing.

    4. RKRay Kurzweil48:31

       Right. Some people feel that we could actually understand what's going on in your brain and actually put things into your brain without actually going into the brain, uh, with something like Neuralink.

    5. JRJoe Rogan48:42

       So something that, like, sits on the outside of your head?

    6. RKRay Kurzweil48:45

       Yeah. Uh, it's not clear to me tha- if that's feasible or not. I've, I've been assuming that you ac- have to actually go in. Now, Neuralink isn't exactly where we want because it's too slow, uh, and it actually will do what it's advertised to do, like if... I actually know some people like this who were active people and they completely lost the ability, uh, to speak and to understand language and so on, um, and so they can't actually say anything to you. Um, and we can use something like Neuralink to actually, uh, have them express something. They could think something and then have it be expressed to you.

    7. JRJoe Rogan49:35

       Right. And they're doing that, right? They had the first patient, the first patient that was-

    8. RKRay Kurzweil49:39

       Yeah.

    9. JRJoe Rogan49:39

       Yeah. And apparently, that person can move a cursor around on a screen.

    10. RKRay Kurzweil49:43

        Right. And therefore, you can do anything. It's, it's fairly slow though.

    11. JRJoe Rogan49:47

        Mm-hmm.

    12. RKRay Kurzweil49:47

        And Neuralink is slow. If you really wanna extend your brain, you, you need to do it at a much faster pace.

    13. JRJoe Rogan49:54

        But isn't that gonna increase exponentially as well?

    14. RKRay Kurzweil49:57

        Yes. Absolutely.

    15. JRJoe Rogan49:58

        So how long do you think it'll be before it's implemented? Where-

    16. RKRay Kurzweil50:02

        Well, it's gotta be by 2045, um, u- um, because that's when the singularity exists and we can actually, uh, multiply our intelligence on the order of a, a million fold.

    17. JRJoe Rogan50:19

        And when you say 2045, w- what is the source of that estimation?

    18. RKRay Kurzweil50:29

        Because we'll be able to, uh... Based actually on, on this chart, uh, and also the increase in, uh, um, the ability of software to also expand, uh, we'll be able to multiply our intelligence a million fold, uh, and we'll be able to, uh, put that inside of our brain. It will be just like it's, uh, part of our brain.

    19. JRJoe Rogan50:58

        So, this is just following the current graph of progress?

    20. RKRay Kurzweil51:01

        Yeah. Exactly.

    21. JRJoe Rogan51:02

        So, if you follow the current graph of progress and if you do understand the exponential growth, then what we're looking at in 2045 is inevitable?

    22. RKRay Kurzweil51:11

        Right.

    23. JRJoe Rogan51:13

        Does that concern you at all or are you excited about it? Uh, do you think it's just a thing that is happening and you're a part of it and you're experiencing it?

    24. RKRay Kurzweil51:22

        I think we'll be enthusiastic about it. Um... I mean, i- imagine if you were to ask a mouse, uh, "Would you like to actually be as intelligent as a human?"

    25. JRJoe Rogan51:38

        Right.

    26. RKRay Kurzweil51:38

        Uh, it's hard to know what people would say, but generally, that's a positive thing.

    27. JRJoe Rogan51:45

        Generally. Yeah.

    28. RKRay Kurzweil51:46

        Um, a- and that's what it's gonna be like. We're gonna be that much smarter.

    29. JRJoe Rogan51:51

        And what do you antici-

    30. RKRay Kurzweil51:52

        A- a- and once we're there, I w- uh, is, is someone gonna say, "No, I don't really like this. I wanna be s- stupid like human beings used to be." Uh, nobody's really gonna say that. D- d- do human beings now say, "Gee, I'm really too smart. I'd really like to be like a mouse."? Uh...
13. 54:03 – 1:01:49

    Perils and power: military use, AGI race, and ‘wrong people’ controlling it

    1. RKRay Kurzweil54:03

       What would really cause a tremendous amount of, uh, danger is something that's not really artificial intelligence, it was invented when I was a child, which is atomic weapons.

    2. JRJoe Rogan54:15

       Right.

    3. RKRay Kurzweil54:16

       Uh, I remember, uh, when I was like five or six, we'd actually go outside, put our hands behind our back-

    4. JRJoe Rogan54:23

       Mm-hmm.

    5. RKRay Kurzweil54:24

       ... uh, to protect us from a nuclear war.

    6. JRJoe Rogan54:26

       Yeah. Drills.

    7. RKRay Kurzweil54:27

       And actually it seemed to work. We're still here, so.

    8. JRJoe Rogan54:30

       (laughs) Do you remember those, uh, things where they tell kids to get under the desk?

    9. RKRay Kurzweil54:34

       Yes. Tha- that's right. We went under the desk and put our...

    10. JRJoe Rogan54:38

        (laughs) Which is hilarious, as if a desk is gonna protect you from a nuclear bomb.

    11. RKRay Kurzweil54:42

        Right. But that's not, uh, AI.

    12. JRJoe Rogan54:45

        Right. No, but AI applied to nuclear weapons makes them significantly more dangerous. And isn't one of the problems with AI is that AI will find a solution to a problem? So, if you have AI running your military, and AI says, "What, you know, what do you want me to do?" And you say, "Well, I'd like to take over Taiwan." And AI says, "Well, this is how to do it." And just implements it with no morals. No, no thought of any sort of, sort of diplomacy or just force.

    13. RKRay Kurzweil55:18

        Right. Hasn't happened yet.

    14. JRJoe Rogan55:20

        Yet.

    15. RKRay Kurzweil55:21

        Because we do have people in charge, and the people are enhanced with AI, and AI can actually help us to avoid that kind of problem, uh, by thinking through the implications of, of different solutions, uh...

    16. JRJoe Rogan55:35

        Sure, if it has some sort of autonomy. But if we get to the point where one superpower has AI, Artificial General Intelligence, and the other one doesn't, how much of a significant advantage would that be?

    17. RKRay Kurzweil55:51

        I mean, I, I do think there are problems. Basically, there's problems with intelligence, and we'd like to say stupid, um, but actually, it's better to be intelligent. Uh, I, I believe it's better to be, to have greater intelligence.

    18. JRJoe Rogan56:11

        Overall, sure. Right. But my question was, if, so if there's a race to achieve AGI, how s- how close is this race? Is it neck and neck? Is it cl- I mean, who's at the lead? And how much capital is being put into these companies that are at the lead? And whoever achieves it first, if that is under con- the control of a government, it's completely dependent upon what are the morals and ethics of that government? What are, what is the constitution? What if it happens in China? What if it happens in Russia? What if it happens somewhere other than the United States? And even if it does happen in the United States, who's controlling it?

    19. RKRay Kurzweil56:49

        I mean, the knowledge of how to create these things is pretty widespread. It's not like-

    20. JRJoe Rogan56:55

        Right.

    21. RKRay Kurzweil56:55

        ... somebody can just, uh, capitalize on a way to do it, and nobody else understands it. Uh, knowledge of how to create a large language model or, uh, how to create the, the, uh, type of chips that would enable you to create this is actually pretty widespread.

    22. JRJoe Rogan57:17

        Mm-hmm.

    23. RKRay Kurzweil57:18

        Um...

    24. JRJoe Rogan57:19

        So, do you think essentially the competition is pretty even-

    25. RKRay Kurzweil57:22

        Yeah.

    26. JRJoe Rogan57:22

        ... in all the countries currently?

    27. RKRay Kurzweil57:24

        Yeah.

    28. JRJoe Rogan57:24

        And there's also probably espionage. There's espionage where they're stealing information and sharing information and selling information, and...

    29. RKRay Kurzweil57:33

        And, uh, in terms of differences, uh, the United States actually has, uh, superior, uh, AI compared to other places.

    30. JRJoe Rogan57:46

        Well, that's good for us.
14. 1:01:49 – 1:51:23

    Engineering humans: identity, emotions, body modification, privacy, and language

    1. RKRay Kurzweil1:01:49

       Right. But, uh, I mean, if we wanna, for example, live indefinitely, this is what we need to do. We, we can't do... We can't-

    2. JRJoe Rogan1:01:58

       What if you're denying yourself heaven? Have you ever thought of that possibility? I know that's a ridiculous abstract concept, but if heaven is real, if the idea of the afterlife is real, and it's, uh, the next level of existence, and you're constantly going through these cycles of life, what if you're stepping in, artificially denying that?

    3. RKRay Kurzweil1:02:17

       That's hard to imagine. I mean-

    4. JRJoe Rogan1:02:19

       It is hard to imagine, but so is life.

    5. RKRay Kurzweil1:02:21

       I-

    6. JRJoe Rogan1:02:21

       So is the universe itself. So is the-

    7. RKRay Kurzweil1:02:23

       Right.

    8. JRJoe Rogan1:02:23

       ... Big Bang.

    9. RKRay Kurzweil1:02:23

       My, my father-

    10. JRJoe Rogan1:02:23

        So is black holes.

    11. RKRay Kurzweil1:02:24

        My father died when I was 22, uh, so it's more than 50, 60 years ago. Um, and, uh, it's hard f- And he was actually a great musician, and he great, created, uh, fantastic music, but he hasn't done that since he died. Um, and there's nothing that exists, uh, that is at all creative, um, based on him. We have his memories. Uh, I actually created a large language model that represented him. I can actually talk to him.

    12. JRJoe Rogan1:03:08

        You do that now?

    13. RKRay Kurzweil1:03:09

        Yeah. Yeah. It's in, it's in the book. Um.

    14. JRJoe Rogan1:03:12

        When you do that, have you thought about implementing some sort of a Sora-type deal where you're talking to him?

    15. RKRay Kurzweil1:03:21

        Well, you can do that now with language.

    16. JRJoe Rogan1:03:23

        Right. But I mean, physically-

    17. RKRay Kurzweil1:03:25

        But we can... We should...

    18. JRJoe Rogan1:03:25

        ... like, looking at him, like you're in a Zoom call with him.

    19. RKRay Kurzweil1:03:29

        Uh, that's a little bit in the future to be able to actually capture the way he looks. Um, but that's also feasible.

    20. JRJoe Rogan1:03:37

        It seems pretty feasible.

    21. RKRay Kurzweil1:03:38

        Yeah.

    22. JRJoe Rogan1:03:39

        And it's... Uh, it certainly... It could be something representative of what he looks based on photographs that you have, right?

    23. RKRay Kurzweil1:03:44

        So, things like that is a reason to continue so that we can create that and create our own ability to, uh, continue to exist. That... You talk to people, and they say, "Well, I don't really want to live past 90 or whatever, 100." Um, but i- in my mind, if you don't exist, uh, there's nothing for you to experience.

    24. JRJoe Rogan1:04:15

        That's true in this dimension. My, my thought on that, people saying that, "I don't want to live past 90," it's like, uh, okay, are you alive now? Do you like being alive now? What's the difference between now and 90? Um, is it just a number, or is it a deterioration of your physical body?

    25. RKRay Kurzweil1:04:30

        Well, eventually.

    26. JRJoe Rogan1:04:30

        And how much, how much effort have you put into mitigating the deterioration of your natural body so that you can enjoy life now?

    27. RKRay Kurzweil1:04:38

        Exactly. And we've actually seen... Who, who would want to take their lives? People do take their lives, uh, if they are experiencing something that's miserable.

    28. JRJoe Rogan1:04:50

        Right.

    29. RKRay Kurzweil1:04:50

        Uh, if they're suffering physically, emotionally, mentally, spiritually, uh, and they just cannot stand the way life, uh, is, is carrying on, then they want to take their lives. Uh, otherwise people don't. Um, if, if they're enjoying their lives, they, they continue. And people say, "Oh, I, I don't want to live past 100." But then when they get to be, uh, you know, 99.9, uh, they don't want to, uh, disappear, uh, unless they're suffering.

    30. JRJoe Rogan1:05:26

        Unless they're suffering. That's what's interesting about the positive aspects of AI. Once we can manipulate human neurochemistry to the point where we figure out what is causing great depression, what is causing anxiety, what is causing a lot of these schizophrenic, uh, people-

Episode duration: 2:03:02