Artificial Intelligence Podcast: Elon Musk: Tesla Autopilot

Full video here.

  • How Elon thought about Autopilot: There are two massive revolutions in the automobile industry. 1) The transition to electrification. 2) Autonomy. It became obvious to me that in the future any car that does not have autonomy would be about as useful as a horse. 

  • We automatically get vast amounts of data because Tesla’s have 8 external cameras, radar, 12 ultrasonic sensors, GPS, and IMU... we are approaching half a million cars that have the full sensor suite. I’m not sure how many other cars have this level of sensors. I’d be surprised if it was larger than 5,000. Which means we have 99% of all the data

  • When someone takes control from Autopilot, it is helpful to think of all input as error. 

  • The thing that is really profound is the hardware currently being produced is capable of full self driving . As we refine the software the capabilities will increase dramatically. Then the reliability will increase dramatically. Then it will receive regulatory approval. 

  • It’s pretty crazy letting people drive a 2-ton death machine manually. In the future people will not believe anyone could drive a car anywhere they wanted. Just like we look at people in the past that were able to move an elevator with a lever however they wanted. 

  • Will current deep learning approaches take us to Artificial General Intelligence? I think we are missing a few key ideas before AGI. But it will be upon us very quickly and then we’ll need to figure out what we should do. If we even get a choice. 

  • A metaphor for the difference between a narrow AI that can navigate streets vs Artificial General Intelligence: Your toaster and your computer are both machines. One is much more sophisticated than another. 

  • Are you confident Tesla can create the world’s best self-driving car? Yes. To me, it seems game, set, match. I don’t want to be complacent or overconfident but this is how appears right now. I could be wrong but it appears Tesla is vastly ahead of everyone

  • When someone creates an AGI and you could ask the AGI one question - what would you ask? What is outside the simulation