Tesla is Going to Win Level 5 – George Hotz  | AI Podcast Clips

Tesla is Going to Win Level 5 – George Hotz | AI Podcast Clips


– [George] Tesla is going to win
Level 5, they really are. – [Lex] So, let’s talk about it. So, you’re focused on
Level 2, currently, currently. – [George] We’re going to be one to
two years behind Tesla, getting to Level 5. – Okay. (laughs) – [George] We’re
Android, right, we’re Android. – You’re Android. – [George] I’m just
saying, once Tesla gets it, we’re one to two years behind. I’m not making any timeline
on when Tesla’s gonna get it. – That’s right, you did. That was brilliant. So, what do you think about
this. Most Level 4 companies are kinda just doing their usual, safety driver, doing
full autonomy kinda testing. And then Tesla does, basically trying to go from lane keeping to full autonomy. What do you think about that approach? How successful would it be? – It’s a ton better
approach, because Tesla is gathering data on a
scale that none of them are. They’re putting real users
behind the wheel of the cars. It’s, I think, the only
strategy that works. The incremental. – Well, so there’s a few
components to Tesla’s approach, that’s more than just incremental. You spoke with… One is software, so, over the air software updates. – [George] Necessity. I mean Waymo and Cruise have those too. Those aren’t. But those differentiate
them from the automakers. – Right, no lane keeping systems have, no cars with lane keeping
systems have that except Tesla. – [George] Yep. And the other one is the
data, the other direction, which is the ability to query the data. I don’t think they’re actually collecting as much data as people think. But the ability to turn on
collection, and turn it off.

31 thoughts on “Tesla is Going to Win Level 5 – George Hotz | AI Podcast Clips

  1. Watch the full episode with George Hotz here: http://bit.ly/2YLIPom If you enjoy it, consider subscribing, sharing, and commenting.

  2. Hotz either does not really understand the problem or he is a con artist. Tesla and everyone else in the autonomous car business is using deep learning or some similar representational model for perception. The problem is that, unlike the brain, a deep neural net is blind to things it has not seen before. This is a fundamental and fatal flaw of deep learning and the reason that the cars are driven millions of miles in the vain hope of capturing all possible situations. How silly is that? George Hotz is out to lunch in my opinion.

  3. His words sound right but I think he’s just a talker, good coder but that’s it. He don’t even have a chip, how can he be 1 yr behind? He’s like the rest of us behind.

  4. I hope Hotz stays true to his words and won't sell out. It would be great to keep having an open source alternative, should really push development forward.

  5. I wonder if Google gives Tesla the IP/GPS data of their customers because I've noticed mine is off in coordinates and always thinks I'm driving when I'm not so they shouldn't include my data in their robots because it might cause them to become unreliable in the future.

  6. Self-driving is a winner take all game. Whoever gets best first, wins. People will buy the most capable product, therefore increasing the data gathering rate, making the system even more robust, causing even more people to buy it….. the perfect positive feedback loop.

    Tesla will have an effective monopoly on Transport, just like Google has an effective monopoly on Search.

  7. I enjoyed watching the full episode just a few days ago. I hope pushing this snippet will spur others to watch it and engage in learning more about self-driving and AI in general.

  8. George was so happy in this interview! You can really see how passionate both him and Lex are about the subject. Them and Elon are the real deal folks!

  9. From the whole video, Mr. Fridman and Mr. Hotz think we're still far enough away from level 5 capability that Tesla must dramatically enhance driver attention monitoring to ensure safe oversight as we transition through Level 2 driving. How long would YOU (Mr. Fridman and all here) guess until we see Level 5? Elon Musk thinks Tesla will turn it on NEXT YEAR!

  10. As it stands, Tesla is not approaching level 5. That's not a threshold on the horizon. They have a part of rather fragile level 2.
    Big data is nice to have but you don't need anywhere near the scale Tesla has access to now. A tiny company could gather plenty in a matter of days so if the data amount would pave the way, Tesla would be way further ahead. Once you have the algorithmic sophistication to be a contender for level 5, the data amount you need is relatively tiny. At the current trajectory, without decisive breakthroughs Tesla is 10-20 years from level 5. Or more. A somewhat full featured level 2 also looks like 5 or more years away. I'm a little surprised at the lack of progress, it does not seem that difficult to me but observing the slow progression over the last 5? years, there is just no way level 5 is happening by end of year nor next year.

  11. Hey Lex this is a great idea! you should put on more clips like this. 5 or 15 minutes segments of the conversations about specific subjects. It would kind of give us more of an idea what the episodes about, and maybe spark extra interest in the podcast overall.

    Kind of like what Joe Rogan does.

  12. If you're going to make a "Clips" thing like Rogan, make a separate channel for snippets and just upload them there.. I'm only interested in the full length vids and don't want to get those smaller ones in my feed.. especially because I've already seen them.. Like this one

Leave a Reply

Your email address will not be published. Required fields are marked *