Sign in
Remember me · Forgot password?

Shit Elon Says - Transcript - Elon Musk Panel BTA 2012

Transcript History  

The reason for SpaceX is I think we need to become a multi-planet species and we were clearly not getting there based on the progress in the space industry and so I started SpaceX to try to solve that problem. That's the overall aspiration. Yeah. If we are not a multi-planet species then the likely probable lifespan of humanity is much less. Plus, it's going to be much more boring, I think. [You're not just trying to service the ISS.] Right, absolutely, that's just the appetizer.

It was very nerve wracking, the first flight to the space station. It was the third flight of our Falcon 9 rocket, so I wasn't perhaps as worried about that, and we had flown our Dragon spacecraft and maneuvered around and reentered, so I knew that the basic spacecraft functionality worked, but we had not tested the proximity operations and berthing system, and there was a bunch of upgrades to the avionics. We went from being mostly single fault tolerant to being - often no fault tolerant to being too fault tolerant. So it's quite a big increase in the complexity of the avionics software. We also added solar arrays and a radiator. Yeah. [You can't test that on Earth.] Right, you can sort of try to get a close approximation but you can't get it exactly what it's like to be in zero-g.

It was a little worrying at first because the LIDARs would not lock on. As we approach the space station, we scan the space station with a thing kind of like a laser radar. It laser scans the space station and it recreates that model internally in the computer to figure out the relative position and movement of the two. We needed a definite lock, as opposed to an ambiguous lock, and the actual space station is a little different from the model of the space station. [It had some highly reflected spots.] It did, yeah. So, essentially, the computer wasn't 100% sure. It was pretty sure, but not 100% sure that it had a lock on the space station. So, what we were able to do is just upload new software to diagnose the problem and then run that software in simulation on a hardware-in-the-loop version of Dragon that we have on the ground. We have a complete version of Dragon, which includes all the avionics software, vales and everything, and we can run accelerated simulations on that system and verify that it would work. Yeah, so we had a proposed code change, ran it on our hardware simulator, it worked, uploaded the code patch, and achieved lock on the space station with that. Essentially we narrowed the field of view. Yeah, essentially we narrowed the field of view like putting blinkers on a horse.

Well, technically astronauts have been on the spacecraft, at the space station. We've already done a lot of work to pass the space station safety review. As Dragon is approaching the space station, it's a robotic space freighter, basically, and if something were to go wrong it could potentially destroy the - uhh.. wrong word, heh.. shudder at the thought - but in an absolute worst case [..] something terrible could happen to the space station and people on-board could be hurt. Obviously there's a lot of caution that's exercised there, both at SpaceX and at NASA, and before we even get close to the space station we've run a million simulations and we're checking it as we get closer, we're verifying that everything looks good, but it is an autonomous vehicle. It's not as though there's someone with a joystick that's steering it. It's just pausing at various points to ask if it has permission to proceed. Then we look at the data, make sure everything is cool, and then give it the okay to go to the next, essentially, way point, but it's doing all of the guidance and control itself.

To add, essentially, the ascent and descent elements of manned spaceflight, as we essentially have - it's 'man rated' if you will, for being around the space station, but not for the ascent and descent phases. In the ascent phase we need to add a launch escape system, so if something were to go wrong with the booster we can escape from the booster and then we've also got to do some upgrades to the rocket to make sure it's as fault tolerant as possible and for those parts which are not as easy to make fault tolerant, that we've tested the living daylights out of it and it'll have safety margins that are higher than a rocket that would carry a satellite. Typically, it's a rough rule of thumb, you aim for about a 25% margin on a satellite mission but a 40% margin on a manned mission - above the expected flight loads. Now, our abort system I'm pretty excited about because this is going to be the first time there will be abort capability all the way to orbit - or, proper abort capability all the way to orbit - because the escape engines are built into the sidewall of the Dragon spacecraft and they use the same propellant that would otherwise be used for on-orbit maneuvering - because you either need to maneuver on-orbit or you need to escape - you don't need to do both. It's like, one or the other. As a result, we avoid having the big solid rocket on the nose and that has some real advantages. That big solid rocket is so heavy that you can't afford to carry it all the way to orbit, so you usually have to discard it about three minutes into the flight. You don't have high acceleration escape capability all the way to orbit like you will with Dragon. It has enough thrust to accelerate away from the rocket, even in at - what's called Max-Q, which is maximum dynamic pressure. So, even when it's going about mach 1.8 or so, and the rocket is going at full force below it, it still has enough thrust to actually get away from the rocket.

Now, being a liquid rocket, we do have an advantage that we can turn it off. Ya know, and it's pretty easy to turn off. In fact, with liquid rockets the tough part is making it stay on. That's what you need to worry about. But for solid rockets, once they're on they tend to stay on. I'm not a big fan of solid rockets for human transportation. Solid rockets are, themselves, quite reliable - particularly if you don't have segments - but you can't turn them off. So, if something goes wrong with the overall system, you're going to have some things wanting to hit you in the butt that are hard to get away from.

Version 1 of Dragon is pretty basic really. I'd classify it as, kinda, level one reentry and landing, which is parachutes to a water landing. It's fairly straightforward and fairly reliable but what we're going to for, call it, version 1.5 is we're going to go parachutes to a land landing. I'm hoping we'll be able to do that next year, but certainly the year or there after. Then, level 2, or maybe even level 3 - maybe level 2 is like wings - but then level 3 would be landing with thrusters. Another advantage of having the escape thrusters built into the escape wall of the Dragon is that, when you come back, you can use some of them to land. So, they're really high precision and they're redundant thrusters - you have eight - and well, if you lost the right four you could lose four - you could lose at least one - they're in pairs, so. We've gone back and forth on how we can make this better, but there's so many constraints in the system, but this is pretty good. You can lose any one of the engines and still be safe, and then we have parachutes as a backup on top of that, so - and the parachutes are redundant. So, there should be - I can't think of a safer way to do it, honestly, than to have redundant thrusters and redundant parachutes for landing. In a best case scenario you can be quite reusable because you can land propulsively - just like they landed on the Moon - and that sets you up for reuse quite well.

Oh, I guess a lot of people aren't aware that of the 46 missions that we have under contract, only 12 are for NASA. So, the great majority of our missions are actually commercial missions. They're satellite launch missions, and that's both for our Falcon 9 rocket and our Falcon Heavy rocket and we're hopeful that Dragon can have some commercial use as well. Ya know, we're talking to some people. There's this guy Bigelow who wants to do a private space station and that could be pretty cool 'cause then we could, ya know, maybe deliver people to his space station, share costs with NASA, ya know, reduce the burden of NASA in that regard. I think really the general approach is, what are the things that we can do to forward the cause of space, and that's what we'll do. Now, some of those things are not going to be the most commercially - ya know, they wouldn't necessarily be profit maximizing - but I don't really care and that's why I'm making sure I maintain control of the company. SpaceX is going to be a viable company - we have some venture investors and I don't think they're too worried about that - but there are different ways to run a company. You can run a company where you're really under the gun with each quarter and you've got to make a lot of short term optimizations which makes it difficult to make big technology advancements because some of those can take years to pay off, so you're they're just at the wrong interval, ya know, the wrong interval to make those changes if you're only doing things on a quarterly basis.

The goal at Tesla has been to change people's thinking with respect to electric cars. People were operating under the illusion that an electric car would have to be, let's say, aesthetically challenged - ya know, low performance, low range, and kinda look like a golf cart. None of these had to be true, but I'd talk to people and say none of these things have to be true, ya know, and you can have a long range car, the physics is pretty obvious - what's the energy density of lithium-ion and what's the energy usage per mile, it's pretty straight forward, and people would be like, oh, no no no, that can't work. I'd be like, where's the error in the calculations here? These are pretty obvious. Amazingly, people would refuse to - they'd either ignore it or say it can't be done - it's ridiculous. So, we just decided to make a car and that's pretty hard to ignore. Whereas discussions or Powerpoint are much easier to ignore. Everything does work on Powerpoint, so there is reason for skepticism there, but having an actual car and one that is fully homologated for use on the roads and meets all the safety standards and everything, was really important. So we did the Roadster and then after the Roadster people said, oh, sure, you can make a small electric sports car, but you couldn't make a real car. Ya know, like a Mercedes or an Audi that has all of the features and capabilities. So then we made the Model-S. Actually, people were quite skeptical. Then we made the Model-S and people were like, oh, you couldn't possibly ramp up production. Yeah, I could ramp up production. It's sort of like a series of issues.

The goal with the Model-S is to show that you can have a long range electric car and with the recent announcement of the supercharger, we're building a nation wide network of really fast chargers, hence the name 'supercharger' which is originally obviously from the gasoline car industry. The idea there is you'll be able to charge your car with the same level of convenience as you'd normally use your gasoline car. If you start a trip that starts at 9am, by noon you want to stop for a bite to eat, go to the restroom and gas up your car, and that's a good 20 to 30 minutes. We were able to figure a way, with some advanced technologies, to have it such that you could stop for half an hour and have three hours driving recharge. So about a six to one ratio. Which is about the convenience inflection point, for most people, for a long range trip.

We made that work, and then we added solar panels to the supercharger stations to address the long tailpipe argument that says, oh, you're just pushing emissions to the power plant. Well, no, because the supercharger stations will actually generate more electricity than the cars use in recharging. For the supercharger, it's basically the same level of convenience during long distance trips. We'll have it nation wide. We already have it throughout California, so you can drive anywhere in California using the supercharger. It'll have enough solar panels to generate electricity back to the grid on an annual basis. That's how we're sizing them, so they'll be slightly energy positive. Then we're also making it free. Well, it's not entirely free. There's obviously the cost of it is built into the cost of the Model-S but the cost of it was so low that we looked at it and said, well, we could charge some small amount for this or we could really make it free for the Model-S.

[A full charge is a couple of dollars.] Right.. well, it's a little more than that, but if you were to stop and put in, say, 150 miles range, it's about four or five bucks. So it's really not much. People drive long distance a lot less than they think they do. "I wanted to have something that is really profoundly better than a gasoline car for driving long distance." Not just try to equal a gasoline car but try to say, well, what could we do which makes an electric car suddenly better than a gasoline car and so that's where free long distance is pretty awesome. You can't do that with a gasoline car. It's way too expensive. The basic tag line is: drive anywhere, for free, on free sunlight, forever.

"I love being a political football." I guess, with respect to Tesla, here are things are with Tesla. Tesla has over 35,000 people and these are high quality jobs. We manufacture, from the ground up, in the US. At least in respect to Tesla, we're a surge in American manufacturing activity. The car's critically acclaimed. We actually export powertrains to the rest of the world because we have Toyota and Mercedes as customers. We export electric powertrains to those guys. We're a net exporter and we've got balance of payments. We're expecting to be break even next month, which is pretty fast after the introduction of the Model-S, I think. Yeah, so, and then if you look at our market cap, the value that the stock market assigns to Tesla is about three billion dollars. If you were to take every company that's failed in every DOE program, it doesn't add up to $3B. If you were to actually look at this from a venture portfolio standpoint - presumably Mitt Romney understands these things, but maybe not - then the net gain is very significant. You can say, well, what value has been added to the American economy, what value has been subtracted from the companies that didn't work out, and the net is very positive. So one would have to say that the DOE has done a very good job.

We did unveil the Model-X, earlier this year, and the Model-X is an SUV that is built on the same platform as the Model-S. It's got a slightly longer wheel base but otherwise it's on the same platform. It's really addressing the SUV and minivan market. It's got a unique innovation which is the double hinged gull wing door on the side. That's never been done before I believe. Certainly hasn't been done in any production car. We call it the Falcon Wing because when both doors are up it looks sort of falcon-like. I think it's the coolest door. As far as doors go, it's pretty cool. The reason it has to be double hinge is that, if you just made it a single hinge gull wing, the arc as it swings out, it swings out too far and then too high, but as a double hinge you can actually do the same where it basically does this movement. So it's actually going, almost straight up and if you can physically fit between the Model-X and another car, then you can open the door. It's actually more convenient than a minivan door, because a minivan door when that opens it actually comes out and slides, so you can't get to the car from the rear, but with the Model-X you can, when the door's open.

[Question about Mars.] I think in the very beginning it may be similar to an Antarctic station. [Then large scale settlement.] Yeah, absolutely. That's the thing that we should be aiming for long-term, which is to create a self-sustaining civilization on Mars. I mean, that's the thing that will ensure that, in the event of a calamity on Earth, civilization continues. The light of consciousness is not extinguished. Those seems like good things to me. So yeah, I think that's what we should strive for. I don't think we need to anything at the Antarctic, that's nice, but on Mars we should be aiming for a real civilization. Which ultimately means taking, at least, tens of thousands of people. Perhaps ultimately millions of people and millions of tons of cargo, because you've gotta recreate the industrial base of Earth. So, to do that, you need really big rockets launching a lot, obviously.

[Question about asteroid mining and refueling in space.] I don't think that's going to be - we shouldn't make that a requirement on the way to Mars but that's a potential optimization - to mine near-Earth asteroids for propellant - but the near-Earth asteroids are not very big. I guess the Moon is also a potential place for propellant depots. [More water being found on the Moon.] Yeah, it's in like, permanently shadowed craters. It's pretty chilly in there, but you could mine the Moon, potentially for water and you could have propellant depots on the Moon. I'd liken it to when the early colonies in the Americas were being established, and early voyages of discovery. You kinda want to go there and then, if it turns out that having way stations makes that trip more efficient over time then people will build those stations. As soon as you've got that destination, you've got the forcing function, then you'll see people do whatever seems sensible to make that better. You definitely don't need to have to mine asteroid resources to get people to Mars. You definitely don't need to do that, so I'd say - since it's incredibly hard to begin with, unless something needs to happen, I would toss it out, and let people do it if it seems to makes sense later when you've got that forcing function of the trips between the planets. I think the things to do is to produce - this is what I think - we need to build really big rockets and I think they should probably use methane, because that's the cheapest fuel, and you can refuel methane on Mars by taking CO2 from the atmosphere, mining water ice, there's a huge amount of water ice on Mars, and you get CH4 and O2 pretty easy, like, basically. That's the way to go, and just make 'em real big and launch 'em a lot and reusable - must be reusable, this is important.

[What do you do on Mars?] I don't know, lots of exploring? I think you'd probably be working on building infrastructure on Mars and exploring all the interesting things. Like, ya know, Valles Marineris makes the Grand Canyon look tiny, it's kind of cool, go down that and check it out. Olympus Mons, it's kind of a shallow gradient but it's the tallest mountain in the solar system. Exploring a new planet, I think, would be pretty interesting, and then building the infrastructure necessary to make life self-sustaining on Mars.

[Question about next generation batteries.] Well, first of all, I think if we had to, we could, just with incremental improvements to lithium-ion, I think we could turn the entire automotive world to pure electric. It gets harder for airplanes. You need a higher energy density for airplanes. But certainly for cars, boats, trains, lithium-ion could do it. There are some modest breakthroughs happening in lithium-ion to, for example, change the inert from carbon to silicon, and there are some lithium-ion chemistrys like lithium-sulfur which actually have really high density. In the 400 to 500 Wh/kg range. Current state of the art of lithium-ion is about 250 Wh/kg. That said, I do think there's potential for a breakthrough in capacitors. In fact, that's what I was going to be studying at Stanford, where I dropped out, was working on high energy density capacitors and leveraging the equipment that's been developed for chip making and photonics to, ya know, create something that's solid state but precise down, essentially at the molecular level and see if that tens of billions of dollars invested in making really tiny circuits could be used to create an ultracapacitor with high energy density. There are some companies - well, there's one company I know in particular, based in Silicon Valley - it's not E-Store, in case you're wondering - that I think has the potential for a breakthrough in that arena. That could be quite revolutionary.

[Question about product philosophy.] Right, well. If you're a newcomer product, it's really not enough to just be as good as the incumbent product, because people are used to what they're used to - people are set in their ways. In order to get people to change, you have to do something that's meaningfully better. Otherwise, the gradient of change - ya know, the change with respect to time - that change is going to happen slowly. If you want it to happen fast, it's got to be obviously better. That's why we've tried so hard with the Model-S to create a car that's obviously better. Adding things like the supercharger - we could have just said, oh, every time you use a supercharger it's going to be $5 and then people would have to do some math to figure out, okay, what does that mean relative to - and it's like, it's free. People understand free. That's really easy to understand.

[Question about motivating young engineers.] I think it is important to be highly adaptive, and I do think physics is a good framework for thinking. I think, it's like, particularly as you're trying to figure out new things, reasoning from first principles is a good way to go, as opposed to reasoning by analogy. It's computationally easier to reason by analogy and if you tried to reason from first principles all the time, you wouldn't be able to get through your day, but when you're trying to do something new and complicated, that is the way to do it because analogies are not necessarily perfect and they're relying on things that have already occurred, so, if you're trying to make something new then it's not a great way to go. What reasoning from first principles really, just, means is boiling something down to the fundamental truths, or what appear to be the fundamental truths, and reasoning up from there, and then having a good feedback loop. So, you're seeing what happens and adjusting accordingly and then, I think, it's important to be particularly alert to negative feedback because people generally, particularly friends, they'll be reluctant to give you negative feedback, because they don't want to hurt you. The reason they're reluctant to do that is because most people are hurt when they get negative feedback. It's not an unreasonable expectation. So, you have to actually coax people to give you negative feedback. Encourage negative feedback and listen to it carefully, and don't react in a bad way when you receive it. That's really important.

[Question about payment systems on Mars.] I think payment systems are pretty easy, particularly if you don't have to integrate with a lot of legacy stuff then payment systems are super easy. That's just like World Of Warcraft, ya know, credits. How many credits do you have in your database? You don't have exchange rates and have to, like, interface with bills and coins and have credit cards and have a federal reserve and all these things, they complicate things. What Paypal really did is de-complicate things, but Paypal would be, like, super-trivial in a new environment.

[Question about going against the grain.] Well, I think that you just need to show that it's physically possible. Step one would be, like, just envelope it with physics and make sure you haven't done something that, like, violates the conservation of energy or momentum or something like that, because then you're almost certainly wrong. Even if someone tells you you're right. Those things are pretty important. Basically, ensure that success is one of the possible outcomes. That's key. Often these things can be illustrated with some very straight forward equations. You don't need to get too complicated. So, if you can sort of walk people through that, that's helpful, but I don't know. It's just amazing. I've literally been in situations where I point out the series of obvious things - like, three child-like equations - literally 11th grade or something like that, and it obviously comes to this conclusion, and yet people are still not convinced, and I'm like, how can you look at that and not be convinced. Let me tell you one of my pet peeves: space solar power. Okay, the stupidest thing ever. If anyone should like space solar power, it should be me. I got a rocket company and a solar company. I should be really on it, ya know. But it's like, super obviously, not going to work because, ya know, if you have solar panels - first of all, it has to be better than having solar panels on Earth, so then you say, okay, solar panel is on-orbit, you get twice the solar energy - assuming that it is out of Earth's shadow - but you've gotta do a double conversion. You've gotta convert it from photon to electron to photon, back to electron. You've got to make this double conversion, so, okay, what's your conversion efficiency? Hmm. All in, you're going to have a real hard time even getting to 50%. [The solar cells are better.] It does not matter, put that cell on Earth then. See, that's the point I'm making. Take any given solar cell, is it better to have it on Earth, or is it better to have it on orbit? What do you get from being in orbit? You get twice as much sun - best case - but you've got to do a conversion. You've got to convert it the energy to photons - well, you have incoming photons that go to electrons, but you - you've gotta do two conversions that you don't have to do on Earth, which is you've got to turn those electrons into photons and turn those photons back into electrons on the ground, and that double conversion is going to get you back to where you started, basically. So why are you bothering sending them to bloody space. "I wish I could just stab that bloody thing through the heart." BTW - electron to photon converters are not free and nor is sending stuff to space. Then it obviously super doesn't work. Case closed. You'd think. You'd think case closed, but no. I guarantee it's gunna come up another ten times. I mean, for the love of God.

[Question about biological effects of long term spaceflight.] I think it's okay. I mean, I think the whole interplanetary human flight thing being a danger to human beings is somewhat overblown because, clearly, we sent people to the Moon, right, and that's deep space and they've lived to quite an old age. We've not seen any premature deaths, really, of people that have gone to the Moon. So really it's just a question of how long can you be in deep space and there's a certain damage rate per day which is then offset by your body's ability to repair that damage. So, going to Mars and doing a six month journey is going to - you're going to have some slight increased risk of cancer but, from what I've seen, sort of a back of the envelope calculation, that increase in cancer is less than if you smoked on the way there. Although smoking is quite bad, I have to say. There is one thing that people should be quite concerned about, which is solar flares. This is often thought about in the wrong way, where people say, oh, you need to have, like, you need to have 20 feet of water or whatever it is to shield against a serious solar storm and they say, oh, you need to have a sphere of water around you, and that sphere of 20 foot water would be ridiculously expensive, or ridiculously heavy, but that's true, but actually, what you'd really do is you'd have a column of water, pointed at the sun and you'd be in front of the column. That'd make much more sense, and then you've gotta have water anyway, so it doesn't end up being a big deal. So I don't think the journey there is - there's no showstoppers there but over time we will find ways to improve it and reduce the risk of cancer and that sort of thing and reduce the journey length as well, but really, fundamentally, we need to get there. If we can't get there, it's all like academic, so we need to get there.

Old New Diff   Stop


TODO: diff