April 14, 2019 by Paul Fosse
In this article, I will highlight the points raised by Elon Musk and Lex Fridman in a recent YouTube video on artificial intelligence and give you the date and timestamp of the various topics in the video. For more context, we also released a preview of the video on Friday.
Lex Fridman is a researcher with a Ph.D. in computer science working on artificial intelligence centered on humans. You know who Elon Musk is. CleanTechnica countless articles on self-driving vehicles and I wrote two myself about Tesla's new neural network chip and its Waymo strategy.
Lex Fridman recently made the headlines by publishing an article concluding that Tesla drivers remained vigilant while remaining vigilant while monitoring their Tesla on the improved autopilot.
- 0:00 – Introduction by Lex
- 3:15 – Elon states that "in the future" Any car that does not have autonomy will be as useful as a horse. "
- 3:50 – Autonomous cars are worth 5 to 10 times more.
- 4:20 am – The car's display is a health check of the car's perception of the outside world.
- 5:40 – Lex asks if Elon will show the uncertainty of the computer. Elon responds that there are two debugging views, one in augmented reality in which boxes are surrounded by objects and one tag of what it thinks is, and another more complex view. They do not think that these debugging views would be useful to consumers.
- 7:45 am Lex asks if he needs to devote efforts to algorithms, data or hardware. Elon states that they have data from radar, 8 cameras and 12 ultrasonic sensors in more than 400,000 cars and that they are only deploying the fast hardware that it took 3 years to develop.
- 10:10: Elon clearly states that "any entry is an error", which means that if the customer has to disengage the autopilot for any reason, it analyzes what needs to be improved.
- 2:50 pm – What Elon thinks is "really deep and we'll emphasize it during the next day of investors" is that cars in production will be able to "drive independently" with just a live update. Of course, Elon has been saying this for almost 3 years and progress toward this goal has been slow at first.
- 3:20 pm – You buy an asset that appreciates.
- 16:10 – Today people are enjoying the autopilot for highway driving. Tesla just needs to improve it so that it works in the surface streets. Cars must recognize lights, complex intersections and complex parking lots.
- 5:45 pm – How long will it take to keep your hands on the wheel? Elon thinks regulators will approve once the probability of accidents, injuries and deaths will be 200% better than that of human drivers. I think that Elon is optimistic. I think the regulators will want the system to be at least 10 times safer than a human, and it will be a major battle over the next few years.
- 8:00 pm – Elon says regulators follow the press and Tesla generates a lot of press. There are 40,000 deaths a year with cars, but every Tesla crash hits the headlines.
- 9:00 pm – Lex explains his study on 18,000 disengagements during which he found that the drivers were attentive.
- 9:40 pm – Elon thinks that human vigilance will quickly become a topic of debate, because the system is so much better than a human, the intrusion of the man can actually harm the security.
- 22:40 – Elon gives the following example: the operators of human lifts were totally replaced by automated elevators. Now, it would be detrimental to safety to allow human intervention in the elevator (beyond pressing the button).
- 24h20 – The rate of improvement of the Tesla neural network is exponential.
- 25:20 – Unlike competitors who are tightly interfering with their products, Tesla allows users to use the autopilot anywhere they can detect lines.
- 26.30 – In the future, people will look back and think that it's crazy to let people drive cars.
- 27:10 – Lex asks Elon about the recent incident in which a car was trapped to make it into the lane with a traffic jam in the opposite direction. Elon said that neural networks are essentially matrix mathematics, that hacking requires very sophisticated knowledge and that his team would easily find it easy to detect matrix hackers.
- 28:20 – People have no idea how neural networks work.
- 28:50 – We miss some ideas to have a general intelligence, but it will be on us very soon.
- 29h50 – Game, set, match – Elon is confident. Tesla is far ahead of the competition in building the best autonomous car in the world, but he admits, "I could be wrong."
- 31:00 – Can a machine love you? Elon declares that if you can not say it, then he can. Perception = reality. Elon channels Keanu Reeves in his hit film "Red Pill or Blue Pill" from the 1999 blockbuster The matrix When Lex asks Elon what he would ask the creator of the simulation, what question would he ask? "What's out of the simulation?"
Go here to watch the video for yourself!
It is clear that Elon has the feeling that he has the advantage by using artificial intelligence to produce autonomous cars. It is also easy to dismiss his assertions of overconfidence, but I have just returned from the return of the launch of Falcon Heavy and I find it useful to consider what Elon has done in this area. Elon is not a rocket scientist, but he trained in economics and physics and learned rockets on his own. He looked at what NASA had done over the past 40 years and said with confidence that he could build a rocket 10 times cheaper than the huge NASA team. After 15 years in the industry, he did it and his goal is now 100 times cheaper instead of 10 times cheaper. Elon is often overly optimistic about the pace of progress, but he is rarely mistaken about his outrageous claims.
If you want to take advantage of my Tesla referral link to get 1,000 miles of free Supercharging on a Tesla Model S, Model X or Model 3, here is the code: https://ts.la/paul92237 (but as I l I said before, if another owner helped you more, please use his link instead of mine).