Good. I just wanted to make sure there was no miscommunication. Thanks.
You made some intelligent and thoughtful points that I wanted to discuss further.
Can we get a summary?
Certainly, there are self-driving developers like Waymo who are way ahead of Tesla in the self-driving department. And Waymo has already deployed self-driving cars at near L4 and is accepting passengers on a regular basis. So, Waymo will most likely deploy full blown driverless robotaxis before Tesla does. But Waymo will focus mostly on small fleets of robotaxis in select cities.
Going along with your idea, I think the specific threat that Tesla should be afraid of is Waymo, MobilEye, Aurora or similar self-driving tech company partnering with an auto maker to actually deploy commercially available, compelling, self-driving cars to the public. Let's face it: the big thing that has kept Tesla ahead is not FSD but the fact that for a long time, Tesla cars have been very compelling cars in other areas. Other auto makers have lagged behind in producing a stylish, compelling electric car. But that is changing. I've been fascinated recently with the Byton M-Byte and K-Byte prototypes and the Lucid Air. They are stylish and compelling cars with excellent FSD hardware. I know Lucid Air plans to use MobilEye. So if these cars can combine their style with the right FSD tech, then you could potentially have very compelling electric cars that also have better FSD than Tesla. Tesla's best hope is really that the partnerships take longer than expected to bear fruit or that these start-ups run into financial or production troubles, to give Tesla more time to finish FSD. Otherwise, I think it is inevitable that at some point auto makers, in partnerships with self-driving tech companies, will start producing and deploying L4 autonomous vehicles. It is just a question of when. In fact, it is probably no surprise that Musk is promising FSD in 2020. If you look at other companies, a lot of them seem to think that 2020-21 is when L4 self-driving cars will happen.
I know Musk has been adamant about the hardware being good enough for FSD but I do secretly hope that Tesla will upgrade the sensor suite at some point in the near future. Honestly, I don't think Tesla really has a choice if they are serious about actually getting to L4/5 autonomy. And yes, I know that if the camera vision is good enough, that you don't really need LIDAR. But I still think that the extra redundancy is beneficial. The issue is not really do you need a sensor as the primary, but what extra reliability can you get with sensor fusion, by having multiple sensors checking the same data. For example, take the stalled car problem. Yes, good enough camera vision can solve that problem. But isn't it better to have the extra redundancy of camera and radar and LIDAR so that you have 3 sensors all checking for that stalled car to increase the probability that your car will not collide with it? Of course it is.
The problem is that Tesla has spent so many years selling "FSD capable" cars that it would be a total betrayal to announce that they are adding LIDAR or other sensors to future cars now. And upgrading the FSD computer post delivery is one thing but Tesla could not afford to upgrade all cars with extra sensors now. I think Tesla's best option now is to do the best they can with the current hardware, try to provide something as close to FSD as possible on the current hardware so that customers get something for their FSD money but then add more sensors on future cars after that. Tesla might have been concerned that the AP2 computer could not handle extra sensor data and that might explain the limited hardware. But with the FSD computer and possible FSD 2.0 computer later, Tesla will have a good computer for FSD that should be able to handle extra sensor data.
So yeah, I am personally hoping that in a few years that Tesla will have a refresh of the Model 3 or S that includes LIDAR, extra radar, extra cameras as well as the FSD 2.0 computer. I think by then the software should be good enough and with the extra hardware, I would be more confident that Tesla actually has L4/5 autonomy. I would be very excited for that. Of course, I have a very personal interest in seeing that happen since it would time perfectly with when I would be ready to trade up my Model 3.
I think this might be the biggest concern for Tesla. As other auto makers start offering more and more compelling electric cars, especially ones that actually offer real FSD, there is the real risk that customers will start abandoning Tesla. For example, if you are an early AP2 owner who is disappointed that Tesla has not kept their FSD promises, you might very well jump ship if you see another auto maker offering a great electric car with real FSD powered by MobilEye or Waymo tech. I have 4-5 years left on my Model 3 auto loan. But once my Model 3 is paid off, I will be watching what other auto makers are offering. By then, Tesla should have some good competition too. I love Tesla. I definitely want my next car to be another Tesla. But in 4-5 years, if another auto maker can offer a great electric car, one that has style, tech, long range and also has real L4/5 autonomy, I might jump ship.
I think the best thing that Tesla can do to combat this problem is simply to press on and try to deliver as many compelling new Autopilot features going forward. There is no better way for Tesla to satisfy customers and pay down that "promise debt" then by actually delivering future autopilot features. Getting new autopilot features should placate owners at least for the time being.
Yes, Tesla's deployment speed is probably Tesla's single biggest asset in FSD. Although, other auto makers are starting to add OTA update capability to their new cars. But Tesla can use OTA updates to continue to add Autopilot features to get it ever closer to FSD. In fact, I think that is Tesla's strategy. As I see things, Tesla's strategy has been to continually add features to its ADAS system to make it more and more complete over time, until it is so advanced that it is indistinguishable from FSD. We see that with the current features on the website. Tesla has gotten Autopilot really good on the highway. Now, they hope to add Navigate on AP on city streets so that Autopilot can handle city streets as well as highways. Then adding more and more functionality, upgrading the NN, Autopilot will eventually be able to handle all driving scenarios.
