马斯克在 TED2022 现场谈 Twitter、Tesla 与大脑运作

0:02Hello.

0:04So, in just a few minutes Elon Musk will be joining us here live on stage for a conversation.

0:14Rumor has it, there are a few things to talk about with him.

0:17We will see.

0:22But before that, I just want to show you something special.

0:29I want you to come with me to Tesla's huge Gigafactory

0:33in Austin, Texas.

0:35So the day before it opened last week, the evening before,

0:38I was allowed to walk around it, no one else there.

0:42And what I saw there was honestly pretty mind-blowing.

0:46This is Elon Musk's famous machine that builds the machine.

0:50And his view, the secret to a sustainable future,

0:53is not just making an electric car,

0:55It's making a system that churns out huge numbers of electric cars with a margin so that they can fund further growth.

1:04When I was there, none of us knew whether Elon would actually be able to make it here today,

1:09so I took the chance to sit down with him and record an epic interview.

1:14And I just want to show you a nine-minute excerpt of that interview.

1:21So here from Austin, Texas, Elon Musk.

1:24Want us to switch now to think a bit about artificial intelligence.

1:27I'm curious about your timelines and how you predict and how come

1:31some things are so amazingly on the money and some art.

1:34So when it comes to predicting sales of Tesla vehicles, for example,

1:39I mean, you kind of been amazing. I think in 2014, when Tesla had sold that year 60,000 cars,

1:46you said 2020, I think we will do half a million a year.

1:49Yeah, we did all this exactly half a million.

1:51Five years ago, last time you came to town, I asked you about full self-driving.

1:56And you said, yep, this very year, I am confident that we will have a car going from

2:03LA to New York without any intervention. Yeah, I don't want to blow your mind, but I'm not always right.

2:12But so talk about what's the difference between those two? Why has full self-driving in particular

2:17have been so hard to predict.

2:20I mean, the thing that really got me

2:21and I think it's going to get a lot of other people

2:23is that there are just so many false storms

2:26with self-driving where you think you've got the problem,

2:32have a handle on the problem, and then it, nope,

2:34turns out you just hit a ceiling.

2:38And because if you were to plot the progress,

2:44the progress looks like a log curve.

2:45So it's like a series of log curves.

2:48So most people don't want to log curve business, but.

2:51It goes up sort of a fairly straight way.

2:57And then it starts tailing off.

2:58And you start.

2:59And there's a kind of ocean.

3:00And diminishing returns.

3:02In retrospect, they seem obvious.

3:03But in order to solve full self-driving properly,

3:07you actually just, you have to solve real world AI.

3:12You know, because you say like,

3:13what are the road networks designed to work with?

3:16The designs work with a biological neural net, our brains,

3:20and with vision, our eyes.

3:24And so in order to make it work with computers,

3:29you basically need to solve real world AI and vision.

3:35because we need cameras and silicon neural nets

3:43in order to have self-driving work for a system

3:46that was designed for eyes and biological neural nets.

3:52I guess when you put it that way,

3:53it's sort of like quite obvious

3:55that the only way to solve for self-driving

3:57is to solve real world AI and sophisticated vision.

4:01What do you feel about the current architecture?

4:02Do you think you have an architecture now

4:04where there is a chance for the logarithmic curve,

4:08not to tell off any time soon.

4:12Well, I mean, admittedly these may be

4:15an infamous last words, but I actually am confident

4:18that we will solve it this year.

4:20That we will exceed,

4:22so you say like the probability of an accident,

4:26what points do you exceed that of the average person?

4:29I think we will exceed that this year.

4:31We could be here talking again in a year.

4:33It's like, well, another year went by and it didn't happen.

4:36But I think this is the year.

4:38Is there an element that you actually deliberately

4:40make aggressive prediction timelines

4:43to drive people to be ambitious?

4:46Without that, nothing gets done.

4:48So it feels like at some point in the last year,

4:52seeing the progress on understanding

4:56that the Tesla AI, understanding the world around it,

5:00led to a kind of an aha moment, it has to.

5:03Because you really surprised people recently when you said,

5:06probably the most important product development going on

5:09at Tesla this year is this robot optimist.

5:13Is it something that happened in the development

5:16of full self-driving that gave you the confidence to say,

5:18you know what, we could do something special here?

5:21Yeah, exactly.

5:22So it took me a while to realize that in order

5:26to solve self-driving, you really needed to solve

5:29real world AI.

5:31And the point of which you saw real world AI for a car,

5:34which is really a robot on four wheels,

5:37you can then generalize that to a robot on legs as well.

5:42The things that are currently missing

5:44are enough intelligence for the robot

5:48to navigate the real world and do useful things

5:51without being explicitly instructed.

5:53So the missing things are basically real world intelligence

5:57and scaling up manufacturing.

6:00Those are two things that Tesla is very good at.

6:02And so then we basically just need to design the specialized

6:08actuators and sensors that are needed for a human-wide robot.

6:11People have no idea.

6:12This is going to be bigger than the car.

6:16But so talk about, I think the first applications

6:19you've mentioned are probably going to be manufacturing.

6:21But eventually the vision is to have these available for

6:24people at home.

6:26If you had a robot that really understood the 3D

6:30architecture of your house and knew where every object in that house was or was supposed

6:38to be and could recognize all those objects. That's kind of amazing, isn't it? The kind

6:44of thing that you could ask a robot to do would be what? Like tidy up.

6:49Yeah, absolutely. Or make dinner, I guess, more the lawn.

6:55Take a cup of tea to grandma and show her family pictures.

6:59Exactly, take care of like grandmother and make sure you're exactly.

7:04And it could recognize, obviously recognize everyone in the home.

7:07Could play catch with your kids.

7:09Yes. I mean, obviously we need to be careful

7:11if this doesn't become a dystopian situation.

7:14Like, I think one of the things that's going to be important

7:17is to have a localized ROM chip on the robot

7:22that cannot be updated over the air,

7:25where if you, for example, would it say stop, stop, stop,

7:27that would, if anyone said that,

7:29then the robot would stop, you know, type of thing.

7:31And that's not updateable remotely.

7:34I think it's gonna be important

7:35to have safety futures like that.

7:37Yeah, that sounds wise.

7:39And I do think they should be a regular career,

7:41since for many years.

7:42I don't love being regulated,

7:44but I think this is an important thing for public safety.

7:47And do you think there will be basically,

7:48like say, say 2050 or whatever,

7:50that like a robot in most homes is what they will be

7:55and people will probably love them and count on them.

7:57You'll have your own butler, basically.

8:00Yeah, you'll have your sort of buddy robot, probably.

8:03Yeah.

8:04I mean, how much of a buddy, do you,

8:05like, do you have any applications you thought is there,

8:08you know, can you have a romantic partner, a sex partner?

8:10I mean, a lot of money out there.

8:11It's probably inevitable.

8:13I mean, I did promise the internet that I'd make catgirls

8:15who could make a robot catgirl.

8:17I mean I'm trying to say that, you know.

8:24So yeah I guess it will be what whatever people want really, you know.

8:30What sort of timeline should we be thinking about of the first models that are actually

8:36made and sold?

8:38Well, you know the first, you know, that we tend to make art for jobs that are dangerous

8:46and things that people don't want to do.

8:48And, you know, I think we'll have, like,

8:50an interesting prototype sometime this year.

8:53We might have something useful next year,

8:55but I think quite likely within at least two years.

8:59And then we'll see rapid growth year over year of the usefulness

9:02of the human-wide robots at decreasing costs and scaling up production.

9:06How are we on the economics of this?

9:08So what do you picture the cost of one of these being?

9:11Well, I think the cost is actually not going to be crazy high.

9:15like less than a car.

9:17Yeah, but think about the economics of this.

9:19If you can replace a $30,000, $40,000 a year worker,

9:26which you have to pay every year,

9:28with a one-time payment of $25,000 for a robot

9:32that can work longer hours, doesn't go on vacation,

9:37and it could be a pretty rapid replacement of certain types of jobs.

9:42How worried should the world be about that?

9:45I wouldn't worry about the sort of putting people out of a job thing.

9:48I think we're actually going to have and already do have a massive shortage of labor.

9:53So I think we will have not people out of work but actually still a shortage of labor even

10:03in the future.

10:04But this really will be a world of abundance.

10:09Any goods and services will be available to anyone who wants them.

10:14It'll be so cheap to have goods and services.

10:16It'll be ridiculous.

10:22So that is part of an epic 80-minute interview,

10:29which we are releasing to people, members of TED 2022,

10:35right after this conference.

10:37You should be able to look at it on the TED Live website.

10:42There's public interest in it.

10:44We're putting that out to the world on Sunday afternoon,

10:47I think, Sunday evening.

10:48but if you're into this kind of stuff,

10:52definitely a good thing to do over the weekend.

10:54Now then, hearing from Elon Live,

10:57there's huge public interest in that.

11:00We have opened up this segment to live stream.

11:03And so we're joined right now by, I think,

11:07quite a few people around the world.

11:09Welcome to Vancouver, welcome to TED22.

11:11You're joining us in the last day of our conference

11:14here in a packed theater.

11:16And we've been hearing all week from people with dreams

11:21about what the next era of humanity is going to be.

11:27And now, arguably, the biggest visionary of them all.

11:30Elon Musk.

11:40Hey.

11:41Elon, welcome.

11:46So, Elon, a few hours ago, you made an offer to buy Twitter.

11:57Why?

11:57How'd you know?

12:04A little bird tweeted in my ear or something.

12:08I don't know.

12:09Have you seen the movie Ted About the Bear?

12:11I have.

12:12It's a good movie.

12:14It's not yet.

12:19So yeah, yeah.

12:21So, was there a question?

12:26Why make that offer?

12:27So, well, I think it's very important for there to be an inclusive arena for free speech where,

12:38oh, yeah.

12:43Twitter has become kind of the de facto town square.

12:47So it's just really important that people have both the reality and the perception that

12:57they're able to speak freely within the bounds of the law.

13:02And so one of the things that I believe Quater should do is open source the algorithm and

13:09make any changes to people's tweets, if they're emphasized or de-emphasized, that action

13:16should be made apparent so anyone can see that action has been taken.

13:21So there's no sort of behind the scenes manipulation either algorithmically or manually.

13:28But last week when we spoke, Elon, I asked you whether you were thinking of taking off.

13:36You said no way.

13:37You said, I do not want to own Twitter.

13:39It is a recipe for misery.

13:41Everyone will blame me for everything.

13:44What on earth changed?

13:45No, I think everyone will still blame me for everything.

13:48Yeah.

13:49If something, if I acquire a Twitter and something goes wrong, it's my fault, 100%.

13:54I think there will be quite a few arrows, yes.

13:57It will be miserable, but you still want to do that.

14:00I mean, I hope it's not too miserable, but I just think it's important to the funk.

14:08It's important to the function of democracy.

14:14It's important to the function of the United States

14:19as a free country and on many other countries

14:21and to help actually to help freedom in the world,

14:25more broadly than the US.

14:26And so I think it's,

14:32I think the risk, the civilizational risk is decreased

14:37if Twitter, the more we can increase the trust of Twitter as a public platform.

14:44And so I do think this will be somewhat painful and I'm not sure that I will actually be able

14:48to acquire it.

14:51And I should also say the intent is to retain as many shareholders as is allowed by the

14:58law in a private company which I think is run 2000 or so.

15:02So it's definitely not from the standpoint of let me figure out how to monopolize it.

15:07maximize my ownership of Twitter, but we'll try to bring along as many shoulders as we

15:13are allowed to.

15:14You don't necessarily want to pay out $40 or whatever it is billion in cash.

15:17You would like them to come with you.

15:20I mean, I could technically afford it.

15:24I have that.

15:26But what I'm saying is this is not a way to sort of make money.

15:35I think this is, it's just that I think this is,

15:38this could, my strong intuitive sense is that

15:43having a public platform that is maximally trusted

15:48and broadly inclusive is extremely important

15:55to the future of civilization.

15:57But you've described yourself.

15:58I don't care about the economics at all.

16:01Okay, that's core to here.

16:03This is not about the economics.

16:04It's for the moral good that you think will achieve.

16:07You've described yourself, Elon, as a free speech absolutist.

16:11But does that mean that there's literally nothing

16:14that people can't say and it's okay?

16:17Well, I think, obviously, Twitter or any forum

16:23is bound by the laws of the country that it operates in.

16:27So, obviously, there are some limitations

16:31on free speech in the US,

16:34And of course, Twitter would have to abide by those rules.

16:39So you can't incite people to violence.

16:42Like a direct incitement to violence.

16:46You can't do the equivalent of crying fire in a movie theater,

16:50for example.

16:51No, that would be a crime.

16:53Right.

16:53It should be a crime.

16:54But here's the challenge.

16:57It's such a nuanced difference between different things.

17:00So there's incitement to violence.

17:04That's a no if it's illegal.

17:06There's hate speech, which some forms of hate speech are fine.

17:11I hate spinach.

17:14Maybe if it's a sauteed and a cream source, it can be quite nice.

17:19But the problem is, so let's say someone says, OK, here's one tweet.

17:24I hate politician X. Next tweet is, I wish politician X wasn't alive,

17:30as some of us have said about Putin right now, for example.

17:33So that's legitimate speech.

17:36Another tweet is, I wish PolyticianX wasn't alive

17:39with a picture of their head, with a gun sight over it.

17:43Or that plus their address.

17:46I mean, at some point, someone has

17:49to make a decision as to which of those is not OK.

17:52Can an algorithm do that?

17:54Or surely you need human judgment at some point?

17:57No, I think the, like I said, in my view,

18:02Twitter should match the laws of the country,

18:07and really there's an obligation to do that.

18:14But going beyond that and having it be unclear

18:19who's making what changes to where,

18:23having tweets sort of mysteriously be promoted

18:25and demoted with no insight into what's going on,

18:29having a black box algorithm promote some things

18:32and other things, I think this can be quite dangerous.

18:36So the idea of opening the algorithm is a huge deal,

18:40and I think many people would welcome that,

18:42of understanding exactly how it's making the decision.

18:45And critique it.

18:47And critique it.

18:47Like I mean, I think the code should be on GitHub, you know?

18:52So people can look through it and say,

18:55I see a problem here.

18:57I don't agree with this.

18:59They can highlight issues.

19:01Suggest changes in the same way that you sort of update Linux or or signal or something like that, you know

19:07But as I understand it

19:10Like at some point right now

19:12What the algorithm would do is it would look at for example how many people have flagged a tweet as obnoxious

19:19and then

19:20At some point a human has to look at it and make make a decision as to does this cross the line or not

19:27But that the algorithm itself can't I don't think yet

19:29tell the difference between legal and okay and definitely obnoxious.

19:34The question is, which humans make that call?

19:39Do you have a picture of that?

19:41Right now, Twitter and Facebook and others,

19:44they've hired thousands of people to try to help make wise decisions.

19:49The trouble is that no one can agree on what is wise.

19:52How do you solve that?

19:53Well, I think we would want to err on this.

19:58If in doubt, let the speech let it exist.

20:03If it's a gray area, I would say let the tweet exist.

20:12But obviously, in a case where there's perhaps a lot of controversy

20:19that you would not want to necessarily promote that tweet.

20:24So I'm not saying this is the, I have all the answers here.

20:29But I do think that we wanna be just very reluctant

20:34to delete things and have just be very cautious

20:39with permanent bans, you know, timeouts I think are better

20:44than permanent bans.

20:47And just in general, like I said,

20:54It won't be perfect, but I think we wanted to really have,

21:00like I said, the perception and reality that speech is as

21:02free as reasonably possible.

21:05And a good sign as to whether there's free speech

21:08is someone you don't like allowed to say something

21:13you don't like.

21:15And if that is the case, then we have free speech.

21:18And it's dev annoying when someone you don't like

21:20says something you don't like.

21:22That is a sign of a healthy, functioning,

21:26free speech situation.

21:31So I think many people would agree with that.

21:33And look at the reaction online.

21:34Many people are excited by you coming in

21:37and the changes you're proposing.

21:38Some of this are absolutely horrified.

21:40Here's how they would say it.

21:41They would say, wait a sec.

21:42We agree that Twitter is an incredibly important town

21:45square.

21:46It is where the world exchanges opinion about life

21:49and death matters.

21:51How on earth could it be owned by the world's richest person?

21:54I can't be right.

21:55So how do you, I mean, what's the response there?

21:58Is there any way that you can distance yourself

22:02from the actual decision making that matters on content

22:05in some very clear way that is convincing to people?

22:10Well, like I said, I think the,

22:13it's very important that the algorithm be open sourced

22:17and that any manual adjustments be identified.

22:21like, so if somebody did something to a tweet,

22:24it's, there's information attached to it

22:27that that action was taken.

22:29And I would personally be, you know,

22:33in their editing tweets.

22:36But you'll know if something was done

22:38to promote demo or otherwise affect a tweet.

22:44You know, as for media sort of ownership,

22:47I mean you've got, you know,

22:49Mark Zuckerberg owning Facebook and Instagram and WhatsApp

22:54and with a share ownership structure that will have

22:58Mark Zuckerberg the 14th still controlling those entities.

23:04So,

23:08like literally.

23:11Well certainly the-

23:12We want to have that at Twitter.

23:13If you commit to opening up the algorithm,

23:16that definitely gives some level of confidence.

23:19So talk about some of the other changes that you've proposed.

23:22So the edit button, that's definitely coming.

23:27If you have your way.

23:28Yeah.

23:29And how do you?

23:30I mean, I think, I mean, frankly, the top priority I have, I would have is eliminating

23:42the spam and scam bots and the bot armies that are on Twitter.

23:51I think these fun influence, they make the product much worse.

24:00If I had a Dogecoin for every crypto scam I saw, we'd have more, about $100 billion Dogecoin.

24:13Do you regret sparking a sort of storm of excitement over Doge and you know where it's gone or?

24:20I mean, I think Doge is fun and you know, I've always said don't bet the farm and Dogecoin

24:29I think I think it's it's I like dogs and I like memes and it's got both of those and

24:37But just on the on the edit button

24:38How do you get around the problem of,

24:40so someone tweets Elon Rocks,

24:42and it's tweeted by two million people.

24:45And then after that they edit it,

24:48so Elon sucks.

24:50And then all those retweets, they're all embarrassed.

24:53And how do you avoid that type of changing of meaning

24:58so that retweaters are exploited?

25:02Well, I think you'd only have the edit capability

25:06for a short period of time.

25:08and probably the thing to do upon the edit would be to zero out or re-tweets and favorites.

25:14Okay.

25:15I'm open to ideas though, you know.

25:19So in one way the algorithm works kind of well for you right now.

25:23I just wanted to show you this.

25:24So this is a typical tweet of mine, kind of lame and wordy and whatever.

25:30And look at the amazing response it gets is this.

25:33Oh my god, 97 likes.

25:37And then I tried another one, and 29,000 lights.

25:47So the algorithm at least seems to be at the moment, you know, if Elon Musk expands to

25:52the world immediately.

25:56Not bad, right?

25:57Yeah, I guess so.

25:59I mean, it's cool.

26:01I mean, you've, but you've, so help us understand how it is you've built this incredible following

26:08on Twitter yourself when, I mean, some of the people who love you the most look at some

26:13of what you tweet and they think it's somewhere between embarrassing and crazy, some of it's

26:20amazing.

26:21I mean, it's a little, but it's, is that actually why it's worked or why is it worked?

26:29I mean, I'm tweeting more or less stream of consciousness.

26:34It's not like, let me think about some grand plan about my Twitter or whatever.

26:38I'm literally on the toilet or something.

26:41I'm like, this is funny.

26:42And then tweet that out.

26:44That's like most of them.

26:47You know, oversharing.

26:50But you are obsessed with getting the most out of every minute of your day.

26:55And so, why not?

26:57You know, so, I don't know, I just like try to tweet out like things that are interesting

27:03or funny or, you know, and then people seem to like it.

27:08So if you are unsuccessful, actually, before I ask that, let me ask this.

27:13In fact, I don't, yeah, so how can I say, is funding secured?

27:24I have sufficient assets to complete the,

27:29it's not a forward looking statement blah blah blah.

27:32But I have to, I mean I can do it if possible.

27:35Right.

27:36So,

27:39and

27:41I'm actually actually even in the,

27:43originally of the,

27:46with Tesla back in the day,

27:48the funding was actually secured.

27:50I want to be clear about that.

27:52In fact, this may be a good opportunity to clarify that.

27:56If funding wasn't being secured, and I should say,

28:01like, why do I not have respect for the SEC in that situation?

28:05And I don't mean to blame everyone at the SEC,

28:08but certainly the San Francisco office.

28:10It's because the SEC knew that funding was secured.

28:15But they pursued an active public investigation, nonetheless.

28:19At the time, Tesla was in a precarious financial situation.

28:23And I was told by the banks that if I did not agree to settle with the SEC, that the

28:28banks would cease providing working capital and Tesla would go bankrupt immediately.

28:32So that's like having a gun to your child's head.

28:36So I was forced to concede to the SEC, unlawfully, those bastards.

28:45Now, it makes it look like I lied when I did not in fact lie.

28:50I was forced to admit that I lied to save Tesla's life, and that's the only reason.

28:55Given what's actually happened to Tesla since then, though, aren't you glad that you didn't take it private?

29:07Yeah, I mean, it's difficult to put yourself in the position at the time.

29:13The Tesla was under the most relentless short seller attack in the history of the stock market.

29:19The best thing called short and distort, where the barrage of negativity that Tesla was experiencing

29:25from short sellers in Wall Street was beyond all belief.

29:29Tesla was the most shorted stock in the history of stock markets.

29:34This is saying something.

29:35So this was affecting our ability to hire people.

29:39It was affecting our ability to sell cars.

29:41was terrible. They wanted Tesla to die so bad they could taste it.

29:51Well, most of them have paid the price. Yes. Where are they now?

30:02So that was a really strong statement. Obviously a lot of people who support you, I would have

30:08you have so much to offer the world on the upside, on the vision side.

30:12Don't waste your time getting distracted by these battles that bring out negativity

30:18and make people feel that you're being defensive or that people don't like fights,

30:22especially with powerful government authorities.

30:25They'd rather buy into your dream.

30:28Aren't you encouraged by people just to edit that temptation out

30:36and go with the bigger story?

30:41Well, I mean, I would say, like, you know,

30:43I'm sort of a mixed bag, you know?

30:45I mean, what?

30:47Well, you're a fighter, and you don't like to lose.

30:56And you are determined that you don't, basically.

30:58I mean, you are.

30:59Sure, I don't like to lose.

31:00I'm not sure many people do.

31:02But the truth matters to me a lot.

31:05that really pathologically it matters to me.

31:09OK, so you don't have to lose.

31:12If in this case you are not successful in the board,

31:17does not accept your offer, you've said you won't go higher.

31:19Is there a plan B?

31:24There is.

31:30I think we would like to hear a little bit about plan B

31:37for another time, I think.

31:39Another time?

31:40All right.

31:44I, that's a nice tease.

31:47All right, so I would love to try to understand

31:54this brain of yours more, Elon.

31:57With your permission, I'd like to just play this.

31:59This is the, oh, actually, before we do that,

32:02he was one of the thousands of questions that people asked.

32:06I thought this was actually quite a good one.

32:08If you could go back in time and change one decision

32:10you made along the way, do your own edit button,

32:13which one would it be and why?

32:15Do you mean like a career decision or something?

32:19Just any decision over the last few years, like your decision to invest in Twitter in

32:23the first place or your anything?

32:28I mean the worst business decision I ever made was not starting Tesla with just JB's

32:37travel.

32:38By far the worst decision I've ever made is not just starting Tesla with JB.

32:43That's number one by far.

32:46All right, so JB Strabble was the visionary co-founder

32:49who was obsessed with and knew so much about batteries.

32:52And your decision to go with Tesla, the companies it was,

32:56meant that you got locked into what you concluded

32:59was a weird architecture now.

33:01There's a lot of confusion.

33:02Tesla did not exist in any,

33:07Tesla was a shell company with no employees,

33:09no intellectual property when I invested.

33:12But a false narrative has been created

33:15by one of the other co-founders, Martin Everhart.

33:18And I don't want to get into the end of the nastiness here,

33:20but I didn't invest in an existing company.

33:24We created a company.

33:27And ultimately, the creation of that company

33:30was done by JV and me.

33:34And unfortunately, there's someone else,

33:37another co-founder who has made it his life's mission

33:41to make it sound like he created the company, which is false.

33:44Wasn't there another issue right at the heart

33:47of the development of the Tesla Model 3,

33:49where Tesla almost went bankrupt?

33:51And I think you have said that part of the reason for that

33:54was that you overestimated the extent to which it was possible

33:58at that time to automate a factory.

34:01Huge amount was spent kind of over automating,

34:04and it didn't work.

34:06And it nearly took the company down.

34:08Is that fair?

34:11I mean, first of all, it's important to understand

34:16like what Tesla actually accomplished

34:19that is most noteworthy.

34:22It is not the creation of an electric vehicle

34:26or creating an electric vehicle prototype

34:29or low volume production of a car.

34:33There have been hundreds of car startups over the years,

34:38hundreds.

34:39And in fact, at one point, Bloomberg counted up

34:42the number of electric vehicle startups,

34:44and I think they got to almost 500.

34:47So the hard part is not creating a prototype

34:49or going into limited production.

34:51The absolutely difficult thing,

34:54which has not been accomplished by an American car company

34:56in 100 years, is reaching volume production

35:00without going bankrupt.

35:02That is the actual hard thing.

35:05The last American company to reach volume production

35:08without going bankrupt was Chrysler in the 20s.

35:12And it nearly happened to Tesla.

35:16Yes, but it's not like, oh geez, I guess if we'd just done

35:19more manual stuff, things would have been fine.

35:21Of course not.

35:23That is definitely not the case.

35:25So we basically messed up almost every aspect

35:31of the Model 3 production line from cells to packs

35:37to drive in voters, motors, body line, the paint shop,

35:44final assembly, everything.

35:49Everything was smised up.

35:51And I lived in the Fremont and Nevada factories for three years,

35:59fixing that production line, running around like a maniac

36:03through every part of that factory, living with the team.

36:10I stepped on the floor so that the team who was going

36:15through a hard time could see me on the floor.

36:19They knew that I was not in some ivory tower.

36:25Whatever pain they experienced, I had it more.

36:29And some people who knew you well, actually thought you were

36:32making a terrible mistake, that you were driving yourself to the edge of sanity almost.

36:37Yes, true.

36:38And that you were in danger of making bad choices.

36:43And in fact, I heard you say last week, Elon, that you, because of Tesla's huge value now

36:49and the significance of every minute that you spend, that you are in danger of sort of obsessing

36:56of it, spending all this time to the point of, to the edge of sanity. That doesn't sound

37:04super wise, isn't there? Like your time, your completely sane, centered, rested time and

37:13decision making is more powerful and compelling than that sort of, I can barely hold my eyes

37:20open. So surely it should be an absolute strategic priority to look after yourself.

37:25I mean, there wasn't any other way to make it work.

37:32There were three years of hell, 17, 18, 18, and 19.

37:39With three years, the longest period of excruciating pain in my life, there wasn't any other way.

37:47And we barely made it.

37:48And we were on the regular edge of the back of the city the entire time.

37:52So when you got drunk.

37:53I don't like it.

37:55Those were three, so much pain.

37:58But it had to be done or tells the word be dead.

38:01When you looked around the Gigafactory that we saw images of

38:05earlier last week and just see where the company has come,

38:09do you feel that this challenge of figuring out the new way of

38:16manufacturing that you actually have an edge now that it's

38:20that you actually have an edge now, that it's different,

38:23that you've figured out how to do this,

38:25and from those three years won't be repeated.

38:30You've actually figured out a new way of manufacturing.

38:36At this point, I think I know more about manufacturing

38:40than anyone currently alive on Earth.

38:45It's me that!

38:46APPLAUSE

38:48Yeah.

38:50I can tell you how every damn part in that car is made.

38:54Basically, you should just live on the factory,

38:55live in the factory for three years.

38:57And...

38:59That was nice.

39:00That was...

39:01For you to know it to something?

39:03Someone wants to compose a symphony to that expression of confidence,

39:07something like that. I have no idea what that is.

39:10Anyway, yeah, every aspect of a car of six weeks this Sunday, I know.

39:15I mean, you talk about scale right now.

39:18You're in the middle of writing your new master plan.

39:21And you've said that scale is at the heart of it.

39:24Why does scale matter?

39:26Why are you obsessed with it?

39:27What are you thinking?

39:28Yeah, well, see, in order to accelerate the advent

39:32of sustainable energy, there must be scale,

39:36because we've got a transition, a vast economy

39:39that is currently overly dependent on fossil fuels

39:42to a sustainable energy economy, one where the energy is,

39:48yeah, I mean, we've got to do it.

39:50So the energy's got to be sustainably generated with wind solar, hydro, geothermal, I'm a believer

40:01in nuclear as well I think ever talk about.

40:06And then since solar and wind is intermittent you have to have stationary storage batteries

40:11and then we're going to transition all transport to electric.

40:17If we do those things we have a sustainable energy future.

40:20The faster we do those things, the less risk we put to the environment.

40:27So sooner is better.

40:30And so scale is very important.

40:34It's not about fresh leases, it's about tonnage.

40:38What was the tonnage of batteries produced and obviously done in a sustainable way.

40:45And our estimate is that approximately 300 terawatt-hours of battery storage is needed

40:52to transition, transport electricity and heating and cooling to fully electric situation.

41:00Others may, there may be some different estimates out there, but our estimate is 300 terawatt-hours.

41:08Yeah.

41:09So we dug into this a lot in the interview that we had put it last week.

41:12And so people can go in and hear that more.

41:13I mean the context is that is I think about a thousand times the current install battery

41:19has I mean the scale up needed is breathtaking basically yeah and and

41:24yeah so so your vision is to commit Tesla to try to deliver on a meaningful

41:30percentage of what is needed yeah and what and call on others to do the rest that

41:34this is what this is a task for humanity to massively scale up our response to

41:39change the energy grade?

41:41Yes, it's like basically how fast can we scale

41:47and encourage others to scale to get to that 300 terawatt hour

41:53installed base of batteries.

41:56Right.

41:57And then of course, there'll be tremendous need

41:59to recycle those batteries, which is,

42:02and it makes sense to recycle them

42:03because the raw materials are like high grade ore.

42:06So if people shouldn't think,

42:07well, they'd be as big pile of batteries,

42:09nothing to get recycled because even a dead battery pack is worth about $1,000.

42:16But this is what's needed for a sustainable energy future.

42:18So we're going to try to take a set of actions that accelerate the day of a sustainable energy

42:25future sooner.

42:26OK.

42:27There's going to be a huge interest in your master plan when you publish that.

42:35Meanwhile, I would love to understand more what goes on in the sprain of yours, because

42:41it is a pretty unique one.

42:43I want to play with your permission this very funny opening from SNL, Saturday Night Live.

42:48Can we have the volume there, actually, please?

42:50Sorry.

42:51It's an honor to be hosting Saturday Night Live.

42:53I mean that.

42:55Sometimes after I say something, I have to say I mean that.

42:59So people really know that I mean it.

43:01That's because I don't always have a lot of international variation in how I speak.

43:07Which I'm told makes for great comedy.

43:11I'm actually making history tonight as the first person with Asperger's to host SNL.

43:22And I think you followed that up with...

43:23At least the first person to admit it.

43:24The first person to admit it.

43:26I mean...

43:27So this was a great thing to say.

43:34But I would love to understand

43:38in how you think of Asperger's.

43:41Like whether you can give us any sense of even you as a boy,

43:44what the experience was,

43:47or as you now understand with the benefit of hindsight.

43:51Can you talk about that a bit?

43:53Well, I think everyone's experience is gonna be

43:56somewhat different.

43:58But I guess for me, the social cues were not intuitive.

44:03So I was just very bookish and I didn't understand this.

44:10I guess others could sort of intuitively understand what does meant by something.

44:20I would just tend to take things very literally as the words as spoken were exactly what they

44:25meant.

44:26But then that

44:29Turn out to be wrong

44:31It's you can't they're not they're not simply saying exactly what they mean. There's all sorts of other things that are meant it took me a while to figure that out

44:38so

44:40You know, we'll eat quite a lot

44:43so

44:46I did I did not have a sort of happy childhood to be frank was quite quite rough and

44:52And I read a lot of books, lots and lots of books, and so that, you know, sort of gradually

45:01sort of understood more from the books that I was reading and watched a lot of movies.

45:07And you know, just, it took me a while to understand things that most people intuitively

45:19understand.

45:20So I've wondered whether it's possible that that was in a strange way an incredible gift

45:26to you and indirectly to many other people.

45:30In as much as brains, you know, are plastic and they go where the action is and if for

45:40some reason the external world and social cues which so many people spend so much time

45:44and energy and mental energy obsessing over, if that is partly cut off, isn't it possible

45:49that that is partly what gave you the ability to understand inwardly the world at a much

45:58deeper level than most people do?

46:01I suppose that's certainly possible.

46:05I think there's maybe some value also from a technology standpoint because I found it

46:12rewarding to spend all night programming computers just by myself.

46:18I think most people, most people don't enjoy typing strange symbols into a computer by themselves all night.

46:24They think that's not fun, but I thought it was, I really liked it.

46:29So I just program all night by myself and I found that to be quite enjoyable.

46:37But I think that is not normal.

46:41I mean, it does, you know, I've thought a lot about, it's a riddle to a lot of people

46:50of how you've done this, how you've repeatedly innovated in these different industries.

46:54And it does, you know, every entrepreneur sees possibility in the future and then acts

46:59to make that real.

47:01It feels to me like you see possibility just more broadly than almost anyone and can connect

47:08So you see scientific possibility based on a deep understanding of physics and knowing

47:14what the fundamental equations are, what the technologies are that are based on that science

47:19and where they could go, you see technological possibility.

47:21And then really unusually, you combine that with economic possibility.

47:26Like what it actually would cost, is there a system you can imagine where you could affordably

47:32make that thing?

47:34And that sometimes you then get conviction that there is an opportunity here, put those

47:39pieces together and you could do something amazing.

47:43Yeah, I think one aspect of whatever condition I had was I was just absolutely obsessed with

47:50truth, just obsessed with truth.

47:56And so the obsession with truth is why I studied physics because physics attempts to understand

48:03the truth of the universe.

48:06Physics is just what are the provable truths of the universe,

48:12and truths that have predictive power.

48:16So for me, physics was a very natural thing to study.

48:21Nobody made me study it.

48:23It was intrinsically interesting to understand

48:25the nature of the universe.

48:27And then computer science or information theory,

48:32also to just understand logic and there's an argument that information theory is actually

48:43operating at a more fundamental level than even physics.

48:49So physics and information theory were really interesting to me.

48:58So when you say truth, I mean, it's not like some people,

49:03so it's what you're talking about is the truth of the universe,

49:05like the fundamental truth that drives the universe.

49:07It's like a deep curiosity about what this universe is,

49:11why we're here, simulation, why not?

49:13We've got time to go into that.

49:15But I mean, you're just deeply curious

49:17about what this is for, what this is, this whole thing.

49:22Yes, I mean, I think the why of things is very important.

49:26I actually, when I was a young teens, I got quite depressed about the meaning of life.

49:37I was trying to understand the meaning of life, looking at reading religious texts and reading books on philosophy.

49:46I got into the German philosophers, which is definitely not wise if you're a young teenager, I have to say.

49:52And we went dark, but dark.

49:54So much better read as an adult.

50:00And then actually I ended up reading the Hitchhiker's Guide to the Galaxy, which is actually a book

50:07on philosophy, just sort of disguised as a silly humor book, but actually it's actually

50:14a philosophy book.

50:16And Adams makes the point that it's actually the question that is harder than the answer.

50:26He sort of makes a joke that the answer is 42.

50:28That number does pop up a lot.

50:31And 420 is just 10 times 42.

50:34Ten times more significant than 42.

50:37Okay.

50:38there's a triangle with 42 degrees and 269s.

50:46So there's no such thing as a perfect triangle, or is there?

50:52But even more important than the answer,

50:56is the question. That was the whole theme of that book.

50:59It's basically how you see meaning then. It's the pursuit of questions.

51:03Yeah, so I have a sort of, you know, a proposal for a worldview or a motivating philosophy,

51:10which is to understand what questions to ask about the answer that is the universe.

51:16And to the degree that we expand the scope and scale of consciousness, biological and

51:22digital, we would be better able to ask these questions, to frame these questions, and to

51:29understand why we're here, how we got here, what the heck is going on.

51:34And so that is my driving philosophy is to expand the scope and scale of consciousness

51:39that we made better understand the nature of the universe.

51:43You know, one of the things that was most touching last week was seeing you hang out with your

51:49kids.

51:50Here's, if I may...

51:52It looks vaguely like a ventriloquist on me there.

51:56I mean, how do you know that's real?

52:02So that's X and you know, it was just a delight seeing you hang out with him.

52:08What's his future going to be?

52:13I mean, I don't mean him personally, but the world he's going to grow up in.

52:17What future do you believe he will grow up in?

52:22Well, I mean a very digital future.

52:25A different world than I grew up in, that's for sure.

52:37But I think we want to obviously do our absolute best to ensure that the future is good for

52:42everyone's children.

52:45And that the future is something that you can look forward to and not feel sad about.

52:52you want to get up in the morning and be excited about the future and we should fight for the

52:56things that make us excited about the future.

53:00The future cannot just be that one miserable thing after another, solving one sad problem

53:06after another.

53:07There have got to be things that get you excited, like you want to live.

53:12These things are very important and you should have more of it.

53:17And it's not as if it's a done deal, like it's all to play for.

53:20Like the future may be horrible.

53:23Still there are scenarios where it is horrible, but you see a pathway to an exciting future,

53:30both on Earth and on Mars and in our minds through artificial intelligence and so forth.

53:36I mean, in your heart of hearts, do you really believe that you are helping deliver that exciting

53:43future for X and for others.

53:48I mean, I'm trying my hardest to do so.

53:53I, you know, I love humanity and I think that we should fight

54:00for a good future for humanity and I think we should

54:02be optimistic about the future and fight to make that

54:05optimistic future happen.

54:13Gailana, I think that's a perfect place to close this.

54:16Thank you so much for spending time coming here and for

54:19the work that you're doing and good luck with finding a wise course through on Twitter

54:24and everything else.

54:25Give them a must.

54:26Thank you.

54:27Thank you.

54:28Thank you.

54:29Thank you.

54:30Thank you.

54:34That was really something.

54:35Thank you, now. That was really something.