Transcript for Elon Musk: Tesla Autopilot

I wouldn't characterize the vision or dream. It's simply that there are obviously two massive revolutions in the automobile industry. one is the transition to electrification and then the other is autonomy. And it became obvious to me that in the future, in any car that does not have autonomy, I would be about as useful as a horse, which is not to say that there's no use, it's just rare. And somewhat, it isn't credit, if somebody has a horse at this point. So, it's obvious that cars will drive themselves completely. It's just a question of time. And if we did not participate in the autonomy revolution, then our cause would not be useful to people relative to cars that are autonomous. I mean, autonomous car is arguably worth 5 to 10 times more than a car which is not autonomous.

Transfer what you mean by long term, but with to at least for the next five years, perhaps ten years.

The whole point of the display is to provide a health check on the vehicle's perception of reality. So the vehicle is taking an information for a bunch of sensors, primarily cameras, but also radar and ultrasonic's GPS and so forth. And then that that information is then rendered into vector space and then you know with a bunch of objects with properties like lane lines and traffic lights and other cars and then in vector space that is re-randered onto a display so you can confirm whether the car knows what's going on or not by looking at the window.

Well, I think it's, I mean, my car, I always look at this sort of the debug view. And there's this two debug views. One is augmented vision, where, which I'm sure you've seen where it basically, we draw boxes and labels around objects that are recognized. And then there's, we'll call the visualizer, which is basically a vector space representation, summing up the input from all sensors. That does not show any pictures, but it shows all of the, it basically shows the cause view of of the world in vector space. But I think this is very difficult for people to not know who people to understand. They would not know what they're looking at.

If like if you're no idea how computer vision works or anything you can still look at the screen and see if the card knows what's going on. And then if you're a development engineer, or if you're, you know, if you have the development build, like I do, then you can see, you know, all the debug information, but those would just be like total diverse to most people.

We automatically get fast amounts of data because all of our cars have eight external facing cameras and radar and usually 12 ultrasonic sensors GPS obviously and I am you and so we basically have a fleet that has and we've got about 400,000 cars on the road that have that level of data actually I think you keep quite close track of it actually yes Yeah, so we were approaching half million cars on the road that have the full sensor suite. So this is, I'm not sure how many other cars on the road have this sensor suite, but I would be surprised if it's more than 5,000, which means that we have 99% of all the data. So there's this huge inflow of data. Absolutely, massive inflow of data. And then it's taken about three years, but now we've finally developed out full self-driving computer, which can process, in order to magnitude as much as the Nvidia system that we currently have in the cars. And it's really just to use it, you unplug the Nvidia computer and plug the Tesla computer in. That's it. And it's, in fact, we're not even, we still are exploring the boundaries of escape abilities. We're able to run the cameras at full frame rate, full resolution, not even crop the images. And it's still got had room, even on one of the systems. The whole sub-driving computer is really two computers, two systems on a chip that are fully redundant. So you could put it both through basically any part of that system and it still works.

I think it's, if you ever say a twin engine aircraft, a commercial aircraft, this system will operate best if both systems are operating, but it's, it's capable of operating safely on one. So, but as is right now, we haven't even hit the edge of performance. So, there's no need to actually distribute to functionality across both SOCs. We can actually just run a full duplicate on on each one.

Well, there's a lot of things that are learned. There's certainly edge cases where I say somebody's on autopilot and they take over and then, okay, that's a trigger that goes to us and I says, okay, they take over for convenience or do they take over because the autopilot wasn't working properly. There's also, like let's say, we're trying to figure out what is the optimal spline for traversing an intersection. Then the ones where there are no interventions are the right ones. So you then say, okay, when it looks like this, do the following. And then you get the optimal spline for a complex, now getting a complex intersection.

Somebody took away some of your sort of manual control from autopilot. And really, like the way to look at this is view, all input is error. If the user had to do input, all input is error.

How do you know if that's going to change with Navi get a notepad which we just released and with that stalk of them so the Navi gauge like lane change based It like a certain control in order to change do it to your lane change or X there freeway or or doing a highway interchange The bathroom journey that will go away with the release that just went out

Yeah, they don't. So if you drive the car, then you do.

It is a huge leap. They also automatically overtakes low cars. So it's both navigation and seeking the fastest lane. So it'll overtake a slow cause and exit the freeway and take highway interchangeers and and then we have traffic light recognition which introduced initially as a warning. I mean on the development version that I'm driving, the car fully stops and goes at traffic lights.

Actually, I don't think we just, the full cell driving computer that we just, that, that has a, or called FSD computer. That's, that's now in production. So if you order any Model S or X or any Model 3 that has the full self-driving package, you'll get the FSD computer. That's important to have enough base computation. Then refining the neural net and the control software. But all of that can be provided as an over there update. The thing that's really profound and well-being emphasizing at the sort of what that investor-day that we're having focused on autonomy is that the cars currently being produced, but the hardware currently being produced, is capable of full self-driving.

the capabilities will increase dramatically, and then the reliability will increase dramatically, and then it will receive regulatory approval. So essentially buying a car today is an investment in the future. You're essentially buying, I think the most profound thing is that if you buy a Tesla today, I believe you are buying and appreciating asset, not appreciating asset.

Yes. A software has no marginal cost, really.

Well, I think people enjoy it very much on the highways. It's a total game changer for quality of life for using Tesla or to pilot on the highways. So it's really just extending that functionality to city streets adding in the traffic light recognition navigating complex intersections. and then being able to navigate complicated parking lots so the car can exit a parking space and come and find you even if it's in a complete maze of a parking lot and then if and then you can just it just drop you off and find a parking spot by itself.

I think it will require Detecting hands on wheel for at least six months or something like that from here. Really, it's a question of like for a regulatory standpoint. How much safer than a person? Just what a pal need to be for it to be okay to not monitor the car. And this is a debate that one can have it. But you need a large amount of data so that you can prove with high confidence, statistically speaking, that the car is dramatically safer than a person. And that adding in the person monitoring does not materially affect the safety. So it might need to be like two or three hundred percent safer than a person. And how do you prove that?

Yeah. Fritch fatalities would be a factor. But there are just not enough fatalities to be statistically significant at scale. But there are enough crashes, you know, there are much formal crashes and there are fatalities. So you can assess where is the probability of crash? than there's another step which probability of injury and probability of opponent injury and probability of death and all of those need to be much better than a person by at least have 200%.

I mean, there's no question that regulators pay disproportionate amount of attention to that which generates press. This is just an objective fact. And it tells of generates a lot of press. So in the United States, there's, I think, almost 40,000 automotive deaths per year. But if there are four in Tesla, they'll probably receive 1,000 times more press than anyone else.

I think this is all really going to be swept. I mean, the systems are proving so much so fast that this is going to be a mood point very soon. Where vigilance is, if something's many times safer than a person, then adding a person just the effect on safety is limited. And in fact, it could be negative.

No, in fact, I think it will become very, very quickly, maybe in towards the end of this year, but I'd say I'd be shocked if it's not next year at the latest, that having the person having a human intervene will decrease safety. I can imagine if you're an elevator. I used to be the elevator operators and you couldn't go in an elevator by yourself and work the lever to move between floors. Now, nobody wants an elevator operator because the automated elevator that stops the floors is much safer than the elevator operator. In fact, it would be quite dangerous to have someone with a lever that can move the elevator between floors.

If you have a system that's at a below a human level reliability, then driver monitoring makes sense. But if your system is dramatically better, more level than a human than driving monitoring, monitoring is not just not help much. And like I said, you just like as you wouldn't want someone into, like you wouldn't want someone in the elevator. If you're in an elevator, do you really want someone with a big lever? So some random person operating the elevator between flows. I wouldn't trust that. I would rather have the buttons.

The right of improvement is exponential.

So anyways, how could it detect lanes with coins?

Well, I mean, I think it's, frankly, it's pretty crazy giving, letting people drive a two-ton death machine manually. That's crazy. Like, in the future, people were like, I can't believe anyone was just a lot to drive for one of these two-ton death machines. I think it's just drive wherever they wanted. It's just like elevators. You're just like, move the elevator with the lever wherever you want. It can stop at halfway between floors if you want. It's pretty crazy. So... It's going to seem like a mad thing in the future that people would drive in cars.

Well, the, you know, when you're on there, it's just like basically a bunch of matrix math. You have to be like a very sophisticated, somebody who really has to have neural nets. And like basically reverse engineer how the matrix is being built. And then create a little thing that's just exactly causes the matrix math to be slightly off. But it's very, then block that by having Basically anti-rate negative recognition. If the system sees something that looks like a matrix hack, excluded. It's such an easy thing to do.

Yeah, like you basically want us to both know what is a car and what is definitely not a car. And you train for this is a car and this is definitely not a car. Those are two different things. I feel like no idea neural nets really. They probably think neural nets involves like, you know, phishing net alarming.

I think we're missing a few key ideas for general intelligence, general artificial general intelligence. But it's going to be upon us very quickly And then we'll need to figure out what shall we do if we even have that choice. But it's amazing how people can differentiate between, say, the narrow AI that allows a card to figure out what a lane line is and navigate streets versus general intelligence. These are just very different things. Like your toaster and your computer are both machines, but once much more sophisticated than another.

The world's best toaster, yes, the world's best self-driving. I'm, yes. To me, right now, this seems game set match. I don't, I mean, that's, I don't want to be complacent or confident, but that's what it is. That is just literally what it, how it appears right now. I could be wrong, but it appears to be the case that Tesla is vastly ahead of everyone.

I think AI will be capable of convincing you to form love with it very well.

You know, we start getting into a metaphysical question of, like, do emotions and thoughts exist in a different realm than the physical. And maybe they do, maybe they don't, I don't know, but from a physics standpoint, I think a tent of think of things, you know, like physics was my main sort of training. And for a physics standpoint, essentially, if it loves you in a way that you can't tell whether it's real or not, it is real. That's a physics view of love. Yeah. If there's no, if you, if you, if you cannot just, if you can't approve that, it does not. If there's no test that you can apply that would make it. make allow you to tell the difference, then there is no difference.

Yes. And there may be ways to test whether it's a simulation. There might be, I'm not saying they're wrong, but you could certainly imagine that a simulation could correct that once an entity in the simulation found a way to detect a simulation, it could either restart that it, you know, pause the simulation, start a new simulation or do one of many other things that then corrects for that error.

What's outside the simulation?