Latest / Elon Musk Podcast / Elon Musk Warns about AI and Future of Humanity

Transcript

Elon Musk Warns about AI and Future of Humanity · Elon Musk Podcast

• 0:00
  This show is brought to you by backblaze.
• 0:02
  I use back place to backup my podcast, my video files, all of
• 0:06
  my writing stuff, and all my photos.
• 0:08
  And you get unlimited computer backup, for Macs and PCs for
• 0:11
  just $7 a month. You can backup your own
• 0:14
  documents photos videos. Drawings projects, all of your
• 0:18
  data and access your backed up data from anywhere in the world
• 0:21
  using the web app and you can access the files on your mobile
• 0:24
  to iOS Android apps all covered. And this is the cool part.
• 0:28
  This is my favorite part. You can restore it by mail.
• 0:30
  A hard drive will come to your house with all your data shipped
• 0:33
  to your door. It could come to your business
• 0:35
  to and you can restore return refund program so you can buy a
• 0:39
  hard drive restore, send the hard drive back within 30 days
• 0:42
  and get a full refund. So basically they ship you this
• 0:45
  hard drive and then you ship it back and you don't ever pay for
• 0:48
  it which is the perfect program for somebody who has huge files.
• 0:51
  You don't want to waste days and days downloading terabytes and
• 0:55
  teraflops of data. And if you're worried about
• 0:57
  accidentally, deleting your files to Bucks extra month, you
• 1:01
  can increase your retention history to one year, and I use
• 1:04
  it for all of my video. Files comes in super handy.
• 1:07
  So seven dollars plus two dollars nine dollars a month and
• 1:10
  you get everything backed up. He's a mind for up to a year.
• 1:13
  And if you use the URL back, place.com slash Ilan, you get a
• 1:19
  fully featured 15-day. No credit card required, free
• 1:23
  trial, check it out, play with it.
• 1:25
  Start protecting yourself from potential, bad times, back your
• 1:28
  stuff up, it's for Recommended by the New York Times ink
• 1:32
  macworld PC World. Life wire wired, Tom's guide and
• 1:36
  I defied mac and more and it's recently been listed on the
• 1:39
  NASDAQ stock exchange under be lze.
• 1:41
  So you know they're legit backblaze is committed more than
• 1:44
  ever to bring an easy and affordable data storage that you
• 1:47
  can trust don't be that person that forgot to back up their
• 1:50
  important files. We've got your back sign up for
• 1:53
  a free 15 day trial no credit card required go there sign up I
• 1:58
  you with it it's we powerful. It is really easy to use so go
• 2:02
  to back place.com slash Ilan back plays.com Ilan
• 2:08
  backblaze.com. He'll on.
• 2:18
  Hello, and welcome back to the most canoes pod, a podcast about
• 2:23
  everything, Elon Musk and in this episode, we're going to be
• 2:26
  talking about AI. So, Elon Musk is a huge
• 2:31
  proponent for, very well thought out AI.
• 2:37
  He also thinks that a, I could be the downfall of humanity.
• 2:40
  You think? Say I could take over Humanity,
• 2:44
  destroy us all. All Terminator situation and
• 2:48
  we're all doomed and he went to an AI conference just recently
• 2:53
  in Shanghai China and talk with Jack, ma the founder of Alibaba.
• 3:01
  This was at the world artificial intelligence conference and well
• 3:07
  I'm just going to play it for you because it's pretty cool.
• 3:10
  And he talks a lot about Harmful AI in this interview, he also
• 3:16
  talks about the good that it can do to humanity and how we should
• 3:21
  kind of become an interplanetary species, a multiplanetary
• 3:24
  species. We need humans on Mars, right?
• 3:26
  That's what he says. So, here we go.
• 3:29
  Here's Elon Musk in Shanghai China at the artificial
• 3:33
  intelligence conference. Yeah.
• 3:41
  I think the, yeah overtime a, I will make jobs kind of
• 3:46
  pointless, probably the last job that will remain will be doing
• 3:50
  writing AI software. And then, eventually they, I
• 3:53
  will just write its own software.
• 3:56
  So, I don't know, I suppose, I would recommend studying
• 3:59
  engineering physics, that kind of thing or working on
• 4:04
  something, where people just want to interact with other
• 4:07
  people, and people enjoy fundamentally interacting with
• 4:11
  other people. So, if you're working on
• 4:14
  something that involves people or engineering, it's probably a
• 4:20
  good approach. You know, art of course.
• 4:28
  Yeah, like I said, I think we're gonna have to figure out this
• 4:34
  neurolink situation. Otherwise, we will be left
• 4:39
  behind. It's very important.
• 4:43
  We do this quickly. Time.
• 4:47
  We don't have much time. We don't have a much time of
• 4:52
  what we don't have much time to solve the neuro-link.
• 4:57
  Yeah. Yeah.
• 4:59
  Is there anything like technology like technology and
• 5:03
  Technology awareness? There's like it's like if it was
• 5:06
  like a topological map of Technology awareness, it's
• 5:10
  mostly flat with a few short buildings and then some very
• 5:15
  tall spires very tall spires and unless you are on that very
• 5:20
  tools fire. It's not obvious what the
• 5:23
  topology is. Yeah, I would say try to learn
• 5:29
  as much as much as possible that allows you to predict the future
• 5:35
  or make the future. So the saying is the best way to
• 5:38
  predict the future is to make it just and then assess, whether
• 5:44
  what you are learning, is enabling you to predict the
• 5:48
  future with less error. All you less wrong.
• 5:52
  We always wrong to some degree, but can you reduce the error on
• 5:58
  your future predictions? I think that's the way to look
• 6:02
  at education as well, of course. But with both creative, create
• 6:08
  the future and predict the future.
• 6:11
  So that includes Arch and all those other things.
• 6:14
  But close the loop on being less wrong about future.
• 6:19
  I would say that's the word right?
• 6:20
  Way to think about education. I mean, down the road with a
• 6:24
  neural link. You can just upload any subject
• 6:28
  instantly. So we like the Matrix, you want
• 6:30
  to fly a helicopter? No problem.
• 6:32
  Well, helicopter will fly themselves, but you know, if you
• 6:35
  wanted to do whatever any given skill, we just uploaded to
• 6:39
  instantly in the way. Education Works right now, it's
• 6:43
  extremely low. Bandwidth is extremely slow.
• 6:48
  Lectures of the worst, really? It was like, very slow.
• 6:56
  All right, I'm going to take a quick break, a little pause for
• 6:58
  the cause and thank my sponsors. You can go to audible.com.
• 7:02
  There's a link in the show notes where you can get two free audio
• 7:05
  books. So thank you, audible for
• 7:08
  sponsoring the show. yeah, just Try to predict the future with
• 7:18
  less error. This is the hot.
• 7:19
  This is very hard as you were saying, I'm not sure, it's
• 7:22
  99.9%, but it's not very good. Generally, a prediction of the
• 7:25
  future, but I think often people don't try it.
• 7:27
  The first thing is try, if you don't try okay, you know, got
• 7:31
  it, you're going to try and then, and then adjust based on
• 7:35
  the error of your flight prior predictions according to the
• 7:39
  size, right? Humans can never create another
• 7:43
  animal that is smart and humans, especially well.
• 7:46
  You have so many smart people. It's impossible to make another
• 7:50
  smart people. I very much disagree with that.
• 7:53
  Okay. Yeah, that's good.
• 7:55
  Yeah, but not with a, I mean, the first thing is we should
• 7:59
  assume is that we are very down and we can do.
• 8:02
  We can definitely make things smarter than ourselves.
• 8:03
  I mean, the don't used to be humans, right?
• 8:07
  So then our earliest civilizations, very primitive.
• 8:13
  We didn't have any technology, really, we're just like, running
• 8:15
  around. Of trying to not get eaten or
• 8:18
  strike Survivor winter. Now, we have like Heating and we
• 8:22
  grow food. This is all new stuff.
• 8:25
  So you know things like obviously gotten way more
• 8:27
  smarter than the past Wastewater.
• 8:30
  So that's going to continue. We are not the last step in
• 8:33
  evolution. So the most important thing,
• 8:36
  like I said, the most important mistake, I see a smart people
• 8:39
  making is assuming that they're smart.
• 8:42
  They're not so well, computers actually are already much
• 8:46
  smarter than people on so many dimensions.
• 8:48
  We just keep moving the goal posts.
• 8:50
  So we used to think, like, for example, being good at chess was
• 8:54
  an example of a smart human and then Kasparov was crushed by
• 8:58
  Deep Blue in 97. That was a long time ago. 22
• 9:03
  years. I mean right now your cell phone
• 9:06
  could crush the world champion of Chess.
• 9:08
  Literally go used to be thought of as something that humans were
• 9:14
  better at than computers. Then Lisa doll was beaten for
• 9:17
  was before 1 by Alpha 0. Then a new version of alpha 0
• 9:22
  also should say. Alphago alphago beat least at
• 9:25
  all for one then is Alpha 0 Alpha, 0 crushed, alphago
• 9:29
  hundred 20. Now it's just pointless because
• 9:33
  it just keeps playing itself. Humans are trying to play a
• 9:37
  computer. Go is like trying to fight Zeus.
• 9:41
  It's a not going to work hopeless.
• 9:44
  We are hopelessly hopelessly inadequate in terms of rendering
• 9:49
  and they basically there's just a smaller and smaller corner of
• 9:54
  what of intellectual Pursuits that humans are better than
• 9:58
  computers and that every year. It gets smaller and smaller and
• 10:03
  and soon we will be far far surpassed in every single way
• 10:07
  guaranteed or civilization. Went those are the two
• 10:11
  possibilities? Okay, well, let me tell you,
• 10:15
  like, might be on the on AI is essentially, the, you can view
• 10:20
  the advancement of AI as solving things, with increasing numbers
• 10:23
  of degrees of, with increasing degrees of freedom.
• 10:26
  So, the thing that the thing with the most most degrees of
• 10:30
  Tim is reality, but a is steadily Advanced solving things
• 10:36
  that have more and more degrees of freedom.
• 10:39
  So, obviously, it's something like like Checkers was very easy
• 10:43
  to solve that that we could solve with classical software,
• 10:49
  classical Computing, not really all that challenging.
• 10:52
  And in fact, there is a complete solution for Checkers.
• 10:55
  Meaning it is literally, literally, every impossible,
• 10:59
  every every Version of checkers is known, and then then there's
• 11:06
  chest, which is also worse. Also, had many, many more
• 11:11
  degrees of freedom than Checkers.
• 11:13
  Many orders of magnitude more than Checkers.
• 11:16
  But still a really, I would say hello order of magnitude lower
• 11:20
  degree of Freedom game. Then there's go, which had many
• 11:23
  orders of magnitude more degrees of freedom, than, than chess.
• 11:29
  So it's really just, Stepping through orders of magnitude of
• 11:32
  degrees of freedom. This is the way to go.
• 11:35
  I think view the advancement of intelligence.
• 11:38
  Yeah. And it's really gonna get to the
• 11:41
  point where it just can completely simulate a person in
• 11:45
  every way possible. Like many people simultaneously,
• 11:48
  in fact, I mean, obviously there's a, there's a strong
• 11:50
  argument, we're in a simulation right now.
• 11:54
  So reminds you that joke of like, you know, if life was a
• 11:58
  video game, what would be The review that's like well, the
• 12:03
  graphics are incredible. As the plot is confusing of the
• 12:09
  respawn takes a long time. That's a video game that's life.
• 12:14
  In fly for the video game respawn.
• 12:15
  Take. It takes 20 years to support
• 12:17
  human being and have them be fully conscious.
• 12:22
  I'm worried about the birth rate in which you alluded to earlier,
• 12:26
  the contrary to like, most people think we have like two,
• 12:30
  Many people on the planet. But actually, this is this is an
• 12:34
  outdated view. This is the assuming that AI is
• 12:38
  fine or something that the AI is, there's benevolent future
• 12:41
  with a I I think the biggest problem, the world will face in
• 12:45
  20 years is population. Collapse collapse.
• 12:49
  I want to emphasize this. The biggest issue in 20 years,
• 12:52
  we'll be population collapse. Not explosion, collapse.
• 12:59
  The it's very easy to see what the world will look like in 20
• 13:02
  years because humans, have a 20-year boot sequence.
• 13:06
  So, like you said okay, well, who is born last year, okay.
• 13:09
  Now, you know what the would look like in 20 years so that
• 13:11
  easy. So reclass seller rating
• 13:12
  collapse and then, the carbonara bottle is like, well, what about
• 13:16
  immigration like from where people, march these people, you
• 13:21
  know, there's no zero people are right now.
• 13:22
  So right now it's the machine planet is only some robots
• 13:24
  there. Well, I think, first of all, I
• 13:27
  think humans will solve environmental sustainability.
• 13:29
  See, I do not mean to suggest complacency or that we just take
• 13:33
  it easy. In fact, this is a
• 13:35
  self-fulfilling one fulfilling prophecy.
• 13:39
  We must we must take immediate and dramatic action and we you
• 13:43
  know, we continue the momentum towards environmental
• 13:47
  sustainability and China is actually the world leader in
• 13:51
  this. In fact, I'm not sure how well
• 13:53
  it is known outside of China. Just how much China is a world
• 13:56
  leader in environmental sustainability.
• 13:58
  It's extremely Impressive. I mean, I think half of all the
• 14:01
  electric cars in the world were made in China last year or
• 14:04
  something like that so you know, I do.
• 14:08
  So I don't mean to suggest suggest complacency but I do
• 14:11
  think humans can and will solve sustainability.
• 14:17
  Thank you so much for listening to the show.
• 14:20
  I really do appreciate it. If you like Elon Musk,
• 14:23
  everything, Elon Musk, all of his companies, all of his ideas,
• 14:28
  make sure to hit that subscribe button because I do this all the
• 14:30
  time. I do this at least three days a
• 14:32
  week, talk about new things SpaceX mainly you know, Tesla
• 14:37
  lately. But we're talking about SpaceX
• 14:39
  soon. We're talking about artificial
• 14:41
  intelligence today and we're going to be talking about a lot
• 14:45
  more musk stuff. In the future.
• 14:49
  So we're going to have some cool guests.
• 14:51
  Hopefully, this is a brand new podcast.
• 14:52
  So I also have space news pod which I cover a lot of space X
• 14:57
  stuff. A lot of space like stuff almost
• 14:59
  every single day. There's a new SpaceX story on
• 15:02
  there. So go check out space news.
• 15:04
  Pod if you haven't already thank you so much for my sponsors and
• 15:08
  thank you for taking the time out of your day, to spend it
• 15:10
  here with me. My name is Will Walden and I
• 15:13
  will see you soon.