Sign in
Remember me · Forgot password?

Shit Elon Says - Transcript - Interview at Code Conference 2016

Transcript History  

** Transcription slowly in progress by kdhilliard. Any assistance appreciated. Please coordinate here or PM me at **

Code Conference 2016

Wednesday, 1 June 2016

Rancho Palos Verdes, CA

Participants: undefined EM: Elon Musk, SpaceX & Tesla undefined KS: Kara Swisher, Recode undefined WM: Walt Mossberg, The Verge

0:00:00 WM: Let me start by saying that we're very glad you're here safe and sound.

0:00:04 EM: Thanks. Sorry for being a little bit late. I flew here with the landing gear down, because there was some kind of landing gear issue.

0:00:11 WM: Landing gear was stuck?

0:00:13 EM: There was some kind of warning light. My pilot said that if they retracted the landing gear it may not go back down again. So, probably best to leave it down.

0:00:21 WM: So this even happens to you.

0:00:24 EM: Yeah. Landing gear problems of various kinds.

0:00:26 KS: Anyway, we're glad you're here. We're happy to wait for you. I mean, we're glad you got here safe.

0:00:33 EM: Thanks for having me. It's great to see you guys.

0:00:34 KS: Thank you for coming. I really appreciate it. You kept you promise, which was nice.

0:00:37 EM: Of course.

0:00:38 KS: I think you were drunk when you promised me. But that's OK. I'll take it.

0:00:40 EM: [Laughs]

0:00:42 KS: So, in the couple of years since you've come you've done some astonishing things in terms of substantive stuff with both of your companies. Let's start talking about space and what you've been doing there. Obviously, you've had some success landing the rocket. You've done a bunch of other things where people thought you weren't going to be successful. Talk a little about the progress you think you've made with SpaceX.

0:01:08 EM: Sure. There are a lot of things where I think I didn't think we'd be successful. Probably the most significant thing is being able to land an orbit-class rocket boost stage and bring it both back to Cape Canaveral, land on land, and be able to land on a drone ship out in the ocean. There is a bit of an education process that's needed to understand orbital dynamics because a lot of people are confused with like why the heck are you landing a rocket on a ship in the ocean. That seems pretty inconvenient. And the reason is because that going up and staying up is actually about velocity horizontal to the Earth's surface. There's a huge difference between space and orbit. It's like space you can think of, say, being the international waters boundary for the Pacific Ocean. Like if you go a hundred miles off shore you're technically out of coastal waters, now you are in the Pacific. So it's like technically you're in the Pacific. But orbit is like circumnavigating the globe. It's a really giant difference.

0:02:25 EM: The reason that things go up and stay up is because you're zooming around the Earth so fast that your outward radial acceleration is equal to the inward acceleration of gravity. So those balance out and you have a net zero gravity. So when you see the Space Station, the thing that is a little counterintuitive is that the Space Station is actually zooming around the Earth at seventeen thousand miles an hour.

0:02:52 WM: Even though it seems like it's just ...

0:02:53 EM: It seems really still, but it's moving really, really fast. To put that into perspective, a bullet from a .45 handgun is just below the speed of sound, so the Space Station is going more than 25 times faster than that. And that's what's needed, actually, to go up and stay up. That's why there's the term escape velocity, not escape altitude. There's no such thing as an escape altitude. There's only escape velocity

0:03:26 WM: You need to be a certain speed to escape the gravity of the Earth.

0:03:29 EM: Yeah. You can think of gravity as kind of a funnel in spacetime. Think about it like a coin funnel. It's very much like that, but it's obviously a sort of four dimensional coin funnel. If you spin a marble or a coin on a coin funnel, when it's far out it spins slowly and as it gets closer it spins faster and faster. If you were to start at the bottom of the coin funnel and wanted to exit, you'd spin it horizontally and it would spin out. And that's really how you get to orbit. The gravity well is like a funnel.

0:04:12 WM: How does that effect why you want to land a ship in the ocean.

0:04:17 EM: Because in order to get to orbit, all that matters is your horizontal velocity. Your altitude doesn't really matter. In fact the force of gravity at the nominal boundary of space, 100 kilometers, is almost exactly the same as it is on the surface of the Earth. It's a few percent lower than on the surface of the Earth. In order to go up and stay up the only thing that matters is how fast you are going horizontal to the Earth's surface. So you have that outward radial acceleration, or think about it like tetherball or something like that. It's really that outward accelerate is the thing that matters.

0:05:00 EM: And so when the rocket is going to orbit, the only reason it's going up is to get out of the thick part of the atmosphere. At high velocity the atmosphere is thick as molasses. And so it goes up very briefly, but if you look at a long exposure of the rockets trajectory, you'll see it goes up but immediately curves over and starts going horizontal. At the point at which the stages separate, those two stages, the primary boost stage which is the most expensive part of the rock, at the point at which that staging occurs can be as high as Mach 10. So it's going away from the launch site at ten times the speed of sound. So in order to get back to the launch site you would have to have enough fuel and oxygen to reverse out that velocity and boost back all the way to the launch site. And you just don't have, the physics of it don't really allow you to have that much, it's not about saving money on fuel or anything, its just physically impossible.

0:06:18 EM: Because another sort of thing about if you're in space is that there is nothing to react against. Like whereas an aircraft can circle very easily because it is reacting against air, in vacuum there is nothing to react against. So the only way to go back the other direction is to apply just as much energy as it took you to go forwards. In fact, twice as much, really, because you've got to zero it out and then you've got to ...

0:06:47 KS: You've got to land elsewhere.

0:06:49 EM: Yeah. So the bottom line is this thing is zinging out to sea at ten times faster than a bullet.

0:06:54 WM: At that point it may well be over the ocean because the ocean covers most of the ...

0:06:59 EM: At the point of separation it's not that far away. It's maybe a hundred kilometers away from the launch site, but it is going like hell away from the launch site. So the only way to really land it is to have it continue on that arc, that ballistic arc, and have it land far out to sea on a ship that is prepositioned to a particular latitude / longitude, very precise, to within about a meter. The rocket will then go from vacuum through rarefied air at hypersonic velocity. So when it's in vacuum, obviously you can't use aerosurfaces, you have to use nitrogen jets to control the attitude and position. And then, as it starts to encounter the air we use grid fins, because grid fins, look sort of like a waffle, they work quite well across a wide regime, from very high velocity hypersonics, though supersonic, transonic, and subsonic. It's hard to have aerosurfaces that work well across that entire regime. So once the aerosurfaces become high it used the four grid fins to sort of control its attitude ...

0:08:26 KS: To land itself.

0:08:27 EM: Yeah, it's controlling pitch, yaw, and roll with the grid fins, and those grid fins will then position it to where it's fairly close to the ship, and then it will light, in this case, three of the nine engines to arrest the velocity and then drop to one engine for precision right before landing.

0:08:52 KS: OK, super hard.

0:08:53 EM: That was maybe bit of a long explanation.

0:08:54 KS: No, that's OK. But what we want to get to is, that's supper fucking hard.


0:08:59 EM: There's a video.

0:09:00 KS: So why is that important? What is that, this moment, that important for you?

0:09:06 EM: In order to reuse the boost stage, which is about 70% of the cost of the rocket ...

0:09:15 KS: What's the cost of that? How much is it?

0:09:18 EM: Well, it's sort of on the order of 30 to 35 million dollars.

0:09:22 KS: Right. So you want to save that.

0:09:24 EM: Yeah. It's like I try to tell my team. Imagine there was a pallet of cash that was plummeting through the atmosphere and it was going to burn up and smash into tiny pieces. Would you try to save it? Probably yes. That sounds like a good idea.


0:09:41 KS: Right. OK.

0:09:44 EM: So yes, so we want to get it back, and that way we don't have to make another one. And I think it's quite tragic if rockets get smashed into tiny pieces and ...

0:09:57 WM: So can I ask a question? We've been going to space for what, 50 years, or something like that. Until you started doing this and Jeff Bezos's company has done it, the government never sort of saved the rockets. They never saved the pallet of cash. Why not? And the Russians didn't either. I mean, what was the deal there?

0:10:20 EM: There was some attempt made to do that with the Space Shuttle, but there was no return ... It's the first time that a rocket booster has returned to launch site from an orbital mission and certainly the first time that there's been a landing on a ship out at sea.

0:10:34 WM: But the regular rockets that weren't designed like planes never tried to do this.

0:10:40 EM: Right. The plane thing is not a good idea in my view. Intuitively it seems like a plane should work, but actually if you consider that every mode of transport has a design that is appropriate to its medium, and if you are in space, wings are not very useful. Because there's no air. And then if you want to go somewhere other than Earth there's also no runways. So these are important considerations.


0:11:22 EM: So that's why when they went to the Moon they used propulsive landing.

0:11:26 WM: Right, but what I'm saying is that when they built the Space Shuttle, it sort of was like a, look like a, kind of bulbous plane.

0:11:31 EM: I think that appealed to Congress. Yeah.

0:11:36 KS: It was cool. Cool.

0:11:38 EM: Yeah, looks like an airplane. OK.


0:11:43 WM: Jeff Bezos was here last night and I asked him what's the difference between what you're doing and what Elon Musk is doing, and he said well, I think we have, I think he used the word like-minded, in the general sense of it, and then he went on to explain some difference. But he talked about, and correct me if I'm misquoting him, but I think he was saying this is all about laying the foundations of being able to do greater things by getting the basic infrastructure, of being able to reuse the rockets, down right. Is that correct? Do you have a similar starting point from him?

0:12:24 EM: I think there's certainly some similarities of opinion. I think that both Jeff and I believe that it's important to the future to be a spacefaring civilization and ultimately be out there among the stars. And I think that's the exciting, inspiring future that, I think certainly people in this room want.

0:12:43 WM: Well, particularly after seeing that the asteroids are going to destroy the planet.

0:12:48 EM: I don't view it as ... I think that when I say multi-planet species, that's what we really want to be. It's not like still being a single-planet species buy moving planets. It's really being a multi-planet species and having civilization and life as we know it extend beyond Earth to the rest of the Solar System, and ultimately to other star systems. That's the future that's exciting and expiring and I that's what, you know, you need things like that to be glad to wake up in the morning. Life can't be just about solving problems. There have to be things that are exciting and inspiring that make you glad to be alive.

0:13:31 KS: So what, in the immediate time frame, what is your goal for SpaceX? Now that you've done this, which is a huge accomplishment, what is the plan for you in the immediate time and then the longer range?

0:13:44 EM: Sure. We plan to refly one of the landed rocket boosters, hopefully in about two or three months, something like that. So that will be an important milestone. The stages are looking quite good. It's a really difficult reentry situation, but they're in good shape. We now have four of them, so we want to start reflying them towards the end of summer. And then hopefully by the end of this year we will be launching Falcon Heavy, which will be the most powerful rocket in the world by more than a factor of two. So Falcon Heavy will be on the order of five million pounds of thrust on liftoff, which is about two thirds the size of the Saturn V.

0:14:44 WM: Oh, really?

0:14:45 EM: Yeah.

0:14:46 WM: That's the rocket which took the astronauts to the Moon.

0:14:49 EM: Right, exactly. In fact, we're launching from the same pad.

0:14:53 WM: The very same pad?

0:14:54 EM: The Apollo 11 pad.

0:14:55 WM: Wow, that's amazing. So, you're hoping to launch that Falcon Heavy by the end of this year?

0:15:00 EM: Yeah, that's our aspiration.

0:15:02 WM: Now that's somewhat of a delay from when you first hoped to launch it, right?

0:15:06 EM: Um, yeah, um, but it's not like we had a lot of pressing customers who wanted us to launch it.


0:15:13 WM: OK.

0:15:15 EM: In fact, the first launch will not have any operational satellites. It will be a demonstration launch. And the first operational flights where customers actually want us to launch it are next year. Whereas there are a lot of customers who want us to launch Falcon 9. So about a quarter of our flights are for NASA, but three quarters are for commercial satellites, broadcast and communications satellites, or science missions for other countries, and there's quite a backlog. We had an issue with a rocket last year, so that put about a six month hold on our schedule, so we're sort of backlogged on our launches, and we're trying to get them out as quickly as we can and service our customers. The launches will take place every two to four weeks, which is quite a high launch cadence.

0:16:11 WM: That's a much faster ca-dance than NASA had, right?

0:16:14 EM: Yeah. It will be more launches than anything else in the world. So more than Russia, more than Europe, more than China by next year, certainly.

0:16:25 KS: And largely to deliver customer's satellites?

0:16:29 EM: Yeah, there's a lot of broadcast and communications satellites that are going to geosynchronous orbit. We'll also be launching the new Iridium constellation. Iridium's got a next generation constellation of satellites. I think sixty or seventy satellites. Quite decent sized satellites. That will be many magnitudes improvement over the current Iridium system. So they will have global broadband. That will be a whole bunch of launches. And then next year we will be flying Dragon version two, which is the one that is capable of taking up to seven astronauts to the Space Station. In fact, Dragon 2 is a propulsive lander as well. It's intended to carry astronauts to the Space Station, but it's also capable of being a general science delivery platform to anywhere in the Solar System.

0:17:35 WM: So where are you going with it?

0:17:36 EM: We're going to send one to Mars in 2018.

0:17:40 WM: Wait, wait, wait. 2018, that's for sure?

0:17:43 EM: Yeah, a couple of years.

0:17:44 KS: A couple years. Now, will you be on that flight?

0:17:47 EM: [Laughs] No.

0:17:48 KS: You have talked about this. You said you want to die on Mars, just not on landing. Right, is that correct?

0:17:54 EM: Well, I think if you're going to choose a place to die, then Mars is probably, your know, not a bad choice.

0:18:01 KS: All right.


0:18:04 EM: It's not some sort of Martian death wish, or something. But, yeah, if you're going to born on Earth, die on Mars, sounds pretty good.


0:18:14 KS: Sending this up to Mars, 2018, when will someone like you get there, from your plans?

0:18:20 EM: Sure. So the 2018 mission would be a Dragon version 2, and that, I wouldn't recommend traveling to Mars in that, because it has the interior volume of a large SUV. The trip for Dragon would be on the order of six months. It's a long time to spend in an SUV, I think. [Laughter] It can be done, can be done, but probably not ideal. And it also doesn't have the capability of getting back to Earth.


0:18:57 WM: That seems more important than the space issue.

0:18:59 EM: Yeah, we'll put that in the fine print.

0:19:03 WM: It's like the side effects in a drug ad. 'Cannot get back to Earth.'

0:19:08 KS: We saw the movie. We saw what happened. He got back.

0:19:14 EM: I actually enjoyed the movie.

0:19:17 KS: Do you think he could have gotten back like that. Was it plausible?

0:19:20 EM: Well, I thought there was some interconnection. It was like 80% scientifically correct. It did connect a series of improbable events.

0:19:29 KS: Such as?

0:19:31 EM: Well, I don't think you can just take off from Mars on an unguided rocket, really, and then prick your finger on a spacesuit and navigate to a spaceship. [Laughter] Not impossible, just extremely unlikely.


0:19:51 WM: But if you're Matt Daemon, maybe?

0:19:53 EM: Maybe. You have some mad skills, for sure.


0:19:57 KS: So when will people like yourself get there? And I assume you will be first in line for that.

0:20:02 EM: Later this year, in September, at the IAC, which is the big world's space conference, industry space conference, I'm going to be presenting the architecture for Mars colonization. So I think what really matters is being able to transport large numbers of people and ultimately millions of tons of cargo to Mars. And that's what's necessary in order to create a self-sustaining, and not only self-sustaining, but a growing city on Mars.

0:20:40 KS: I'm curious. Have you been to space yet?

0:20:42 EM: No.

0:20:43 KS: Why? I mean you could just go up, right, for a little bit, or not?

0:20:46 EM: I could, I suppose, yeah.

0:20:48 KS: Why haven't you? Like walked around the ...

0:20:49 EM: I think I will at some point. Yeah, I would. Yeah, I probably will, as well.

0:20:52 EM: Will you do a Moon test before you go to Mars?

0:20:57 EM: Um, yeah, I'm probably going to go to orbit in four or five years, or something like that. But again, space and orbit are very different things.

0:21:03 WM: But on the Mars thing, would you send up two or three, whether it's you or not, ... I kind of would prefer if you tried it, frankly, because it would be exciting ... but would you send up some people before you do this whole architecture for colonizing Mars? Just a handful of people to see what ...

0:21:26 EM: The basic game plan is we're going to send a mission to Mars for every Mars opportunity from 2018 onwards. They occur approximately every 26 months. We're establishing cargo flights to Mars, that people can depend on ... for cargo. And, it's like I said, the Earth-Mars orbital rendezvous is only every 26 months, so there's one in 2018, there'll be another one in 2020. And, I think if things go according to plan we should be able to ... [pause] ... we should be able to launch people probably in 2024 with arrival in 2025.

0:22:13 KS: That's soon.

0:22:14 WM: Is that a more certain schedule than United Airlines?


0:22:20 EM: Well, um, I don't know.

0:22:23 KS: Well, if he hasn't been on United Airlines in years ...

0:22:24 EM: There are certainly some uncertainties associated with that. So anyway, that's the game plan, approximately 2024 to launch the first of the Mars Colonial Transport System with people. This will be a very big rocket.

0:22:44 WM: Very big? Bigger than a Saturn V?

0:22:45 EM: Very big, yes.

0:22:46 WM: Twice as big, or what?

0:22:49 EM: September, I'll tell you?

0:22:51 WM: Not going to say anything until September? Come on!

0:22:53 EM: Very big.

0:22:54 WM: Come on!

0:22:55 EM: It has to be very big.

0:22:56 WM: How big is very big?

0:22:58 EM: EM: So big ...


0:23:04 WM: Do you think we should abandon the Earth, at some point?

0:23:07 EM: No! No. No, I think it's great.


0:23:09 WM: But you have said things ...

0:23:12 EM: Why would be abandon Earth? It's really nice here.

0:23:14 EM: You have said things about, we may have to abandon the Earth so it's good to have a plan B. You've said that before.

0:23:19 EM: No I haven't!

0:23:20 KS: No, that was Bezos.

0:23:21 EM: That's, I think that's, maybe Jeff, I don't know, but it wasn't me. [Laughs] It Wasn't Me, like Shaggy.

0:23:30 KS: So let's move to things on this Earth. Let's move to Hyperloop, Tesla, other things. But let's talk about Tesla first.

0:23:35 EM: Sure.

0:23:37 KS: Where do you feel like the company is at, at this point? And there's been lots of activity in self-driving cars, in autonomous. How do you look how everybody's jumped in? Google, Apple, others. And all the car manufacturers.

0:23:54 EM: There have been so many announcements of autonomous EV startups, I'm waiting for my mom to start one. It's like, 'Mom, you too?' [Laughter] I mean, there's a lot. In the US alone, there are four, I think, maybe five, China funded EV startups. At the billion dollar plus level. Like, serious funding. And there's a bunch of startups, and then of course the car industry as a whole seems to be moving in that direction. Volkswagen just announced a huge battery factory they are going to build. And I think these are all good. It's good for the industry to be moving towards sustainable transport as soon as possible. We open sourced our patents to try to be helpful in that regard. It's encouraging to see all this activity. From Tesla's standpoint, we want to take the set of actions that are likely to accelerate the advent of sustainable energy. So, scale up production as fast as we can. So we accelerated plans for the Model 3 by two years. And so we want to get to try to get to half a million cars in total in kind of the 2018 time frame, which is an aggressive schedule, but, I think, achievable. And then maybe a million car a year by 2020. I can see a pretty clear path to get there. Autonomy is obviously extremely important. People are going to want autonomy. It's going to be odd to have a car without autonomy in the future.

0:25:56 KS: How do you look at all these efforts, not your mother ...

0:26:00 EM: Yeah, my mom's, yeah ...

0:26:02 KS: She's not going to do it. She may do a rocket situation ... But how do you look at each of them. Let's go through them. What Google is doing, how do you asses what they are doing, when you are looking at it? Because they will be competitors at some point. These are all eventual competitors.

Google has done a great job of showing the potential of autonomous transport. But they're not a car company, so they would potentially license their technology to other car companies. And I think they announced something with the Fiat, and so I wouldn%u2019t say that Google's a competitor because they're not a car company. They would compete with somebody perhaps with the licensed technology, too, but not to them directly.


Uhm, yeah that will be more direct.

[That will be more direct?]


[You can tell that by the hiring pattern, that kind of stuff?] Yeah. [So how do you assess it?]

I would say that I think it%u2019s great that they%u2019re doing this, and I hope it works out.

[What%u2019s the time frame for them, do you think?]

I don't know; I mean, I think they should have embarked on this project sooner, actually. But I don't know -- they don't share with me the details of their recent plans, but I don't think they'll be in volume production sooner than maybe 2020 -- that's going to be the soonest.

[And is that too late? You say they should have embarked sooner; is that because 2020 would be too late to stop you, or beat you, or compete with you, or what?]

No, it%u2019s just like it%u2019s a missed opportunity. [It'll be over by 2020?] It's that in a couple years, I think they will probably make a good car and probably be successful. The car industry is very big, so it's not that one company leads to the exclusion of others. There are a dozen car companies in the world that are of significance, and the most that any company has is approximately 10% market share. So it's not like somebody comes up with a car and they're suddenly like going to kill everyone else -- it's not that way. And the sheer/share scale of automotive manufacturing is hard to appreciate until you see the plants; they're gigantic. [Yeah, I have seen the plants.] Yeah, the share/sheer size of the industrial infrastructure is staggering.

[Not just the assembly plant, but everything else that goes?]

Yeah, the supply chain is actually only the tip of the iceberg, really. The surface plant is literally tip of the iceberg. The supply chain is, once you go to tier 2, tier 3, tier 4 suppliers, that's probably an order matted to more than the final assembly.

[So you think Google would not be a competitor, Apple probably will be a direct competitor.] Yeah, yeah sure. [What about the car companies?]

Yeah, I think they will be competitors, yeah sure.

[Who do you see out there that has done a nice job so far? Mercedes, or GM -- of a competitive car? A potentially competitive car, I guess.]

I mean, I don't think any of the car companies thus far have made a really great electric car. I mean, tell me if you disagree, but I don't think yet that any of them made a great electric car. They, presumably, will continue to improve on what they've done so far, and then at some point they might make a car that's a great car, but they haven't done that yet.

[Can I ask you about batteries for a second?]

Uh, yeah sure.

[So you%u2019re building this giga factory, right? You -- it's built? It's --]

Well, it%u2019s not completely built.

[Okay, but a part of it's up and running.]

A big track was built, yeah. It%u2019s a really gigantic thing. Like, when the giga factory is done it will be the largest footprint building of any kind in the world, of any kind -- not just factories, literally of any kind.

[What is this? The largest rocket, the largest building. I mean --]

[laughter in the audience]

Well, I mean it%u2019s not scaled for scale's sake. It's just like, if you say, 'Well, we're going to accomplish these goals,' then you kind of have to -- we make a big thing.

[Okay, you've got this big thing; it's this big, giant building.] Yeah. [It's going to make batteries. The batteries it's going to make are lithium --]

Because we%u2019re going to have an opening -- well, it's technically not an opening party, since it's been operating for a little while, but we're going to have a party soon. Do you guys maybe want to come?

[Okay. With them? Sure, we%u2019ll come to the battery party.] Yeah, yeah sure. [Can they -- they can't come? [pointing to the audience] Like, they can't come, right? Just us?]

It%u2019s pretty -- this is what I%u2019m saying -- this is, this is crazy. This is like an alien dreadnought. It%u2019s really nutty.

[Because I love the battery party. But talk about where it%u2019s going.]

[But are these lithium ion batteries?] Yeah, yeah that's true. [So they're the same batteries that's in our phone] No, no. [No? Explain, please explain.]

Yeah, so, lithium ion -- [Wait, have you made a battery break-through or something I'm interested in?]

Uhm, yeah. I mean, generally, there are so much nonsense out there about batteries. Like, you can believe about one percent of what you read -- maybe. Lithium ion covers a very broad range of technologies, and you can have an enormous difference in the power density and the energy density, and the cycle life between one chemistry and another. They can be really, enormously different. Uhm, so, what you really, actually want to ask is what is the cathode and what is the anode?


So, in our case -- [That%u2019s right. [laughs] Okay. I just put it in the place --]

Well, the lithium is two percent of the cell mass. So it's like salt in the salad; it's a very small amount of the cell mass and a fairly small amount of the cost. But it sounds like it's big because it's called 'lithium ion,' but really, our battery should be called 'nickel graphite,' because it's mostly nickel and graphite.

[Again, and --] It%u2019s nickel, cobalt, aluminum -- [But battery --] got little things, and graphite with a silicon-oxide layer on the end of --

[-- battery efficiency, or the power you can store in a certain mass seems to be moved very slowly, at least compared to where we're used to Moore's law, pushing integrated circuits faster, batteries kind of are always in our consumer devices, so it's lagging behind. In your -- you've built this giant thing, the biggest building in the world that anyone's ever seen -- ] hopefully it will, but it's [-- your building --] yeah, yeah [-- the big giant one that anyone has ever seen to make batteries. Your whole business depends on batteries in these cars. Have you figured out a way to do some significant increase in yield of energy from a given amount of space in the battery?]

Well, yeah. I mean, the energy density is increasing sort of maybe on the order of like, five-ish percent per year. And it doesn't sound like much, but you add that up over a number of years with compound interest, it ends up being quite a significant number. And a lot of people sort of think that, oh, we just sort of cobbled together some laptop batteries and somehow made a great car, but if it was that easy, then I think we would have quite a few competitors who did the same thing. But it's really quite a lot harder than that. It is the cylindrical form factor, but the internal battery is quite different from what you would find in a laptop, and it will be increasingly different from what's built at the giga factory, which is highly optimized for automotive and has improved energy density. But mostly, it's not the energy density that is the issue. Because, if you buy a Model S today, the range is around 300 miles, and that's quite a lot. So it's pretty rare that people really need to go more than 300 miles at a time without stopping. [Right.] So I don't think we really have a range issue. We can make a 400-mile ranged car today, but that wouldn't be too big of a deal. What really matters is decreasing the cost per unit of energy of the battery packs so you can make the car affordable; that's actually the important thing. So there's really two main dimensions along which cost optimization and making something available in the national market can be achieved. One is design iteration, going through multiple versions of something, and then the other is economies of scale. You kind of need both of those things in order to make compelling mass market product. If you look at, like, cell phones and how many design iterations have we gone through with cell phones and look at the scale at which they are made, which is enormous; and that's what enables everyone to have a supercomputer in their pocket.

[So, speaking of the sales, while we're talking about the sales, you have booked how many orders for --]

0:44:16 KS: And the despot would be the computer?

0:44:18 EM: Or the people controlling the computer.

0:44:21 WM: And do you worry specifically about any of these companies I mentioned who've all seemed to now be pivoting towards this battleground in the next 10 years?

0:44:34 EM: I won't name a name but there's only one.

0:44:38 WM: There's only one you're worried about? And they're not preoccupied with making a car that will compete with you, I assume?

0:44:46 EM: There's only one.

0:44:52 KS: This is an interesting..

0:44:54 EM: It's not about competing, this is sort of like, what would be the point of competing for mutual destruction, it's not about competing it's really just about trying to increase the probability that the future will be good, that's all, so the goal of openAI is really just to take the set of actions that are most likely to improve the positive futures. Like, if you can think of like the future as a set of probability streams that branch out and then converge; collapse down to a particular event and then branch out again and there's a certain set of probabilities associated with the future being positive and a different type flavours of that and at openAI we want to try to guide.. do whatever we can to increase the probability of the good futures happening. That's really what we're trying to do there.

0:45:45 WM: Do you worry that by making this open, some bad actors may use some of what has been developed to do bad stuff with the power of AI?

0:45:56 EM: Yeah, that is certainly the, i mean, a good rebuttal to that, however, i think if AI power is widely distributed then, and there's not, say, one entity that has some super AI that is a million times smarter than anthing else, if instead the AI power is broadly distributed and, to the degree that we can link AI power to each individual's will, like you'd have your AI agent, everyone would have their AI agent, and then if somebody did try and do something really terrible then the collective will of others could overcome that bad actor, which you can't do if there's one AI that's a million times better than everyone else

0:46:48 WM: And it's proprietary

0:46:50 EM: and it's.. yeah, it either has its own will or more likely, at least in the beginning, is controlled by, you know, some small set of people. So, I think that's really the risk. You know, like I've said, what's the best form of government, big fan of, I think it was Churchill, 'Democracy is the worst form of Government, except for all the others'

0:47:14 WM: Right

0:47:15 KS: So speaking of that, this election, you are from,

0:47:19 EM: Oh, no.

0:47:20 KS: No no no. Yes, yes, yes. How does that strike you what's happening now? You've come to this country, you're naturalised now

0:47:28 EM: I think I'm glad that the framers of the constitution saw fit that the President was someone who was captain of a large ship with a small rudder

0:47:40 KS: OK, And?

0:47:42 EM: There's a limit to how much harm any given President can

0:47:45 KS: Are you sure about that?

0:47:46 EM: Oh yeah, yeah

0:47:47 KS: So you're not worried about.. are you backing either of the candidates at this points?

0:47:52 EM: Try to stay out of this situation

0:47:55 KS: Because?

0:47:56 EM: I don't think that it's the finest moment in our democracy

0:47:59 WM: Well, given it's not the finest moment in our democracy, do you think the best thing is to stay out? or to get in and try to head off the worst?

0:47:05 EM: We'll see. I mean, I'm not sure how much influence I can have, as one person, on the outcome, I mean, if I think I could make a difference I would probably try to do something. Like I said, I'm just glad.. being the US President is like being captain of a large ship with a small rudder, so there's just a limit to how much good or bad the President can do. Obviously if the President could make the economy great, and there was like a button he could press, he'd be pressing that button at the speed of light, but they can't. They can't just magically make the economy good, no President wants the economy bad - ever - there's just a limit to how much they can do. I guess there is the nuclear thing..

0:49:08 WM: Yeah, the nuclear thing. I guess there's the nuclear thing.

0:49:13 EM: Yeah but I don't think we would just arbitrarily launch nuclear missiles

0:49:20 KS: One would hope not

0:49:21 WM: The President can just do that

0:49:22 EM: I don't think so arbitrarily

0:49:23 WM: The President is the commander in chief. No, he's the commander in chief

0:49:27 EM: I still don't think that means you can launch nuclear missiles whenever you want. I think Congress would be quite upset about that and

0:49:35 WM: They might not be consulted

0:49:36 EM: Yeah, but I think that the military would be like 'We really think Congress should be consulted on it before you launch a pre-emptive nuclear strike'.

0:49:43 WM: That might happen, that could happen. Are you willing.. you're basing your faith on that though?

0:49:47 EM: I'm quite confident that the military would not just randomly agree to launching nuclear missiles at somebody

0:49:56 KS: Well that's calming. We're gonna put up just very quickly a ... on Hyperloop. you've been involved in it, your level of involvement is what at this point

0:50:06 EM: Again, I know it's a bit confusing because I

0:50:12 WM: You talked about it when you were here last time..

0:50:14 EM: Yeah. I actually came up with the idea. I came up with an initial idea which turned out to be wrong and it wouldn't work, several years ago. But I sort of shot my mount off and said I have an idea that turned out didn't work but I've, with a lot of iteration, was able to come up with something where the physics hangs together and then published the paper and just said 'Look, anyone who wants to do this, Great, be my guest', Cos I have my plate full running Tesla and SpaceX. So I think it'd be great to have any interesting new transport solutions, anything that gets people to their destination in a way that's safer, costs less, is more convenient. That'd be great. And probably the most valuable thing that the Hyperloop paper that I published has done is to spur thinking in terms of new transportation systems. So it's not just 'Oh lets have a fast train that's not even as fast as what Japan's done in the 80's' - I don't see what the point in that is. We should really be trying to think of something that's, particularly in California, that's Hey, what is the best, what's invented something new that's way better than anything else. 0:51:50 KS: Do you wanna shoot your mouth off about that?

0:51:55:EM Well, I'm not an investor in any of the companies that are working on it. I've tried to be neutral cos I'm not trying to favour one company over another, but just to encourage anyone that is interested; to say try to give them moral support, and I hope they succeed. The only thing that I am doing on the Hyperloop front, we're holding a student competition and the student competition is really just aimed towards encouraging students to think about exciting new transport methods. And it's totally cool if they wanna do some architecture that's different than what is proposed in Hyperloop, and in fact the winning team at the student competition that we held earlier this year used a different suspension mechanism than what I proposed. I proposed using, essentially taking air that essentially builds up on the nose from the compressor and floating that through air skis so that you simultaneously remove the drag from the nose and provide a means of suspending the pod and that's also something that works well even at supersonic velocities. You can go, it's been demonstrated up to Mach 1.1 in terms of using air bearings ** added by anotherother 27/7/16 **

1:14:00 ??: There's a sort of a philosophic concept that a sufficiently advanced civilization would be able to create a simulation

1:14:15 EM: A simulation?

1:14:16 ??: Yeah, maybe you've answered this before?

1:14:20 EM: I've had so many simulation discussions it's crazy. In fact, it got to the point where basically every conversation was the A.I./Simulation conversation and my brother and I finally agreed that we would ban such conversations if we were ever in a hot tub. Because that really kills the magic.

1:14:40 ??: But we're not in a hot tub, so the idea's right, any advanced civilization would create, could create, a civilization that's like our existence and so the theory follows that maybe we%u2019re in the simulation. Have you thought about this?

1:14:56 EM: A lot. So much so it had to be banned from hot tubs. It's not the sexiest conversation ever.

1:15:08 ??: ... and, okay, are we in one?

1:15:10 EM: I think here%u2019s the strongest argument for us being in a simulation- probably being in a simulation, i think is the following: that 40, call it 40 years ago, we had Pong. Two rectangles and a dot. That was what games were. Now, 40 years later, we have photorealistic, 3D, simulations with millions of people playing simultaneously, and it%u2019s getting better every year. And soon we%u2019ll have virtual reality, have augmented reality. If you assume any rate of improvement at all then the games will become indistinguishable from reality. They%u2019ll be indistinguishable. Even if that rate of advancement drops by a thousand from what it is right now. Then you just say okay let's imagine it's ten thousand years in the future, which is nothing in the evolutionary scale. So given that we're clearly on a trajectory to have games that are indistinguishable from reality and those games could be played on any set-top box or on a PC or whatever, and there would probably be, you know, billions of such computers or set-top boxes it would seem to follow that the odds that we%u2019re in base reality is one in billions. Tell me what%u2019s wrong with that argument.

1:16:42 ??: So, is the answer yes?

1:16:44 EM: The argument is probably. But is there a flaw in that argument?%u201D

1:16:52 ??: No, the argument makes sense. So, the assumption then is that somebody beat us to it, and this is a game?

1:16:58 EM: No, there's a one in billions chance that this is base reality.

1:17:01 ??: Oh, okay. What do you think?

1:17:05 EM: Well I think it's one in billions%u2026 I mean that seems to be clearly, what it suggests. And actually, arguably we should hope that that's true because otherwise if civilization stops advancing then that may be due to some calamitous event that erases civilization. So, maybe we should be hopeful that this is a simulation because otherwise%u2026

1:17:34 WM: So they could reboot it?

1:17:35 EM: ...well otherwise either we're gonna create simulations that are indistinguishable from reality or civilization will cease to exist. Those are the two options.

** Hey kdhilliard, I just added his viral bit about simulations, but the majority of the interview is still missing. Hope the formatting looks okay. **

Old New Diff   Stop


TODO: diff