3 ms·

I just don't get how Tesla is not ahead on Self Driving. Waymo has spent 10 years mapping the same city and its basically the single easiest place to do self-dr

by nickik 6y ago

I just don't get how Tesla is not ahead on Self Driving. Waymo has spent 10 years mapping the same city and its basically the single easiest place to do self-driving in the whole world. If you actually look at Tesla presentation, a huge part of the problem they are dealing with is handing things like 50000 different kinds of speed limit signs in every country in the world. Assigning lights to the associated lanes and things like that. Finding the 0.00001% cases like object falling of trucks on the highway. Understanding complex intersection without prior mapping data of any kind. I have not heard a single clear explanation how Waymo will go from where the currently are, to overriding a global general solution. Tesla is a silicon Vally tech company that is trying to get its hand on the exact same people that Apple and Google are. And they unlike Apple and Google have their own gigantic data-gathering fleet across every first-world country. They are so far ahead of everybody in terms of data and data gathering its not even remotely close. The reason Tesla seems to get no respect is that (a) Elon Musk over-promised and now he is forever labeled a joke that can never be taken seriously (b) the consumer product on the road does not yet do all the things it claims it will one day do so it providing something for people to criticize.

screye 6y ago
Tesla is the only company that believes that 2d vision can solve the whole problem. Every other company uses it with LIDAR. It may be hubris, but I simply cannot see how vision alone will be sufficient to solve self-driving. This is especially true in heavy rains, snow and any driving condition that doesn't follow the nicest of layouts like the Bay Area. This means that occlusion will stay unsolved by definition. Even if it works out one day, my intuition is that 3D maps / LIDAR self-driving is going to be ready sooner than that, purely due to having richer data. Also, from what I know, (very little but) Tesla isn't THE company landing top self-driving and vision talent. They seem to be going to Waymo, Uber, Apple and Nvidia instead. Additionally, while I deeply respect Andrej Karpathy, he simply doesn't have the same research credentials as some of the distinguished professors who lead the other groups. 3D self-driving labs are all building on each other's work, as the cost of GPUs and LIDAR sensors has dropped by orders of magnitude over the last decade. Cameras and sensors on the other hand remain more or less where they were 10 years ago, and Tesla is left alone to compete with everyone else.
- nickik 6y ago
  > This is especially true in heavy rains, snow and any driving condition that doesn't follow the nicest of layouts like the Bay Area. CommaAI also believes lidar are a bad idea. Tesla has cars using their technology everyday all over the world and thus have 10000x more insight into how conditions effects their hardware/software. While Waymo has spend 10 years testing in Arizona and basically never once had to contend with these problems. And lidar does have its issues in certain conditions as well, its not some magic technology. > They seem to be going to Waymo, Uber, Apple and Nvidia instead. Very hard to judge that. I would like to see the numbers on that. The main evidence to me is that SpaceX/Tesla always rank highest on 'Where you want to work' for engineering students. Also, it seems to me that Tesla is focusing very having a basic framework that non-super AI experts can improve the system. That is the only way you can even begin to manage the complexity of the global road network something people like Waymo has defiantly not yet even begin to solve. > 3D self-driving labs are all building on each other's work, as the cost of GPUs and LIDAR sensors has dropped by orders of magnitude over the last decade. And Tesla makes its own hardware that dropped even more compared to what everybody else uses. > Cameras and sensors on the other hand remain more or less where they were 10 years ago, and Tesla is left alone to compete with everyone else. Lidar by itself does nothing, you still need cameras and sensors as well, so it is purely added cost. There is simply no denying that high quality cheap lidar that can be deployed on millions of cars doesn't exist yet. I would not be against Tesla using lidar. My more important point is that I think HD mapping is a horrible idea, lidar or not. And as long as that is the way you approach the problem you will never have a general self driving solution.
- rasz 6y ago
  I wouldnt mind if Tesla used Stereo-vision, but they try doing self driving with single view crappy cameras. First it was 30 fps 6640x480 monochrome Mobileye EyeQ3, now its 3? center mounted 36 fps 1280x960 monochrome/red cameras.

by nickik 6y ago

screye 6y ago

Tesla is the only company that believes that 2d vision can solve the whole problem. Every other company uses it with LIDAR. It may be hubris, but I simply cannot see how vision alone will be sufficient to solve self-driving. This is especially true in heavy rains, snow and any driving condition that doesn't follow the nicest of layouts like the Bay Area. This means that occlusion will stay unsolved by definition. Even if it works out one day, my intuition is that 3D maps / LIDAR self-driving is going to be ready sooner than that, purely due to having richer data. Also, from what I know, (very little but) Tesla isn't THE company landing top self-driving and vision talent. They seem to be going to Waymo, Uber, Apple and Nvidia instead. Additionally, while I deeply respect Andrej Karpathy, he simply doesn't have the same research credentials as some of the distinguished professors who lead the other groups. 3D self-driving labs are all building on each other's work, as the cost of GPUs and LIDAR sensors has dropped by orders of magnitude over the last decade. Cameras and sensors on the other hand remain more or less where they were 10 years ago, and Tesla is left alone to compete with everyone else.

nickik 6y ago
> This is especially true in heavy rains, snow and any driving condition that doesn't follow the nicest of layouts like the Bay Area. CommaAI also believes lidar are a bad idea. Tesla has cars using their technology everyday all over the world and thus have 10000x more insight into how conditions effects their hardware/software. While Waymo has spend 10 years testing in Arizona and basically never once had to contend with these problems. And lidar does have its issues in certain conditions as well, its not some magic technology. > They seem to be going to Waymo, Uber, Apple and Nvidia instead. Very hard to judge that. I would like to see the numbers on that. The main evidence to me is that SpaceX/Tesla always rank highest on 'Where you want to work' for engineering students. Also, it seems to me that Tesla is focusing very having a basic framework that non-super AI experts can improve the system. That is the only way you can even begin to manage the complexity of the global road network something people like Waymo has defiantly not yet even begin to solve. > 3D self-driving labs are all building on each other's work, as the cost of GPUs and LIDAR sensors has dropped by orders of magnitude over the last decade. And Tesla makes its own hardware that dropped even more compared to what everybody else uses. > Cameras and sensors on the other hand remain more or less where they were 10 years ago, and Tesla is left alone to compete with everyone else. Lidar by itself does nothing, you still need cameras and sensors as well, so it is purely added cost. There is simply no denying that high quality cheap lidar that can be deployed on millions of cars doesn't exist yet. I would not be against Tesla using lidar. My more important point is that I think HD mapping is a horrible idea, lidar or not. And as long as that is the way you approach the problem you will never have a general self driving solution.
rasz 6y ago
I wouldnt mind if Tesla used Stereo-vision, but they try doing self driving with single view crappy cameras. First it was 30 fps 6640x480 monochrome Mobileye EyeQ3, now its 3? center mounted 36 fps 1280x960 monochrome/red cameras.