5 ms·

Is anyone else confused on the merits of Waymo's argument? They differentiate "driver assistance" from "fully autonomous" systems in that Tesla somehow can neve

by spsful 6y ago

Is anyone else confused on the merits of Waymo's argument? They differentiate "driver assistance" from "fully autonomous" systems in that Tesla somehow can never achieve the latter because they have developed the former. The only specific reason for this that I could find was a mention of LiDAR and how Waymo sees it as indispensable. > Krafcik says that Waymo has largely completed technical work on its self-driving software and is now focused on scaling the technology up. ...cool? Again, no mention of a SPECIFIC reason Tesla cannot do the same work as Waymo outside of not having LiDAR sensors. They attack Tesla for not developing a fully autonomous vehicle from the get-go, but don't seem to explain why this makes it any harder for them to make the finished product.

rcxdude 6y ago
The core point is you can't go smoothly from a driver assist to fully autonomous. They're just completely different from the point of view of the reliability of the software you need. A huge amount of things with driver assist you can write off as 'the driver will take over'. There's not a specific part of you can point to and say 'that's what will stop them', so much as the entire design of the system needs a different level of consideration because it becomes so safety critical, as opposed to safety-enhancing (autonomous driving is probably the most complex safety critical design ever attempted, by at least an order of magnitude. Driver assist is way easier by comparison). It affects the attributes of the sensors, the design of the neural networks interpreting those sensors, the high level design of the decision making of the whole vehicle, the low-level design of the electronics, basically everything. Tesla hasn't put forward a plan which indicates their full autonomous development is going to be anything other than a continuation of their driver assist work (and it's certainly frequently and incorrectly stated that the state of their driver assist puts them ahead in the race to full autonomy), and if they want to be credible to those familiar with the challenges involved, they need to actually show something more than that.
- Udik 6y ago
  > you can't go smoothly from a driver assist to fully autonomous Ok. So Waymo has developed a 100% reliable full self driving system that handles every possible situation. It was never put in a situation in which it would have given back control to a human driver, because that would have meant "going smoothly" to driver assist mode, and we have already said that this is impossible.
  - rcxdude 6y ago
    I was not referring to the dynamic shift while in use, I was referring to the development process. Regardless, the waymo system does not have a driver assist mode, it's either in control or it isn't (and during development it may be overridden by or automatically fall back to a safety driver, who has a job akin to a driving instructor, something which general requires more qualifications and attentiveness than a driver).
jariel 6y ago
The argument is that they are not even in the same category of technology complexity. Which is very reasonable. Tesla's real-world deployment of autopilot is frankly probably an advantage of sorts. And maybe starting with those realities is 'good path'. But it doesn't change the nature of the fact the gap is quite big.
mdorazio 6y ago
A full discussion on this would require an article several times as long as this one as well as some understanding of how vehicle control systems work. I do a fair amount of work with OEMs in this space and to summarize it in a horribly short way: 1) A system that relies primarily on cameras is necessarily limited by camera constraints. No, just because humans primarily rely on vision does not mean an ML model with cameras can do the same and achieve the same results in situations as complex as urban driving. 2) Every major player other than Tesla in this space spends a huge amount of time architecting an entire computing, data exchange, and control platform for autonomous vehicles that is bulky by necessity, but far more robust both in terms of sensor capabilities and ability to cross-check between data inputs to decide what control signals to send. Tesla basically added some cameras and low-resolution radar (they're improving the radar now, but it sure as hell wasn't in the platform they originally promised was FSD-ready) and is hoping it will be enough when everyone else studied the capabilities years ago and decided it wasn't. To summarize, what Tesla has built is a platform with capabilities that very strongly seem to be limited to Level 3 automation, while engineers working on Level 4 long ago concluded that the Tesla approach is insufficient, and then adjusted their own hardware and software strategies accordingly.

by spsful 6y ago

rcxdude 6y ago

The core point is you can't go smoothly from a driver assist to fully autonomous. They're just completely different from the point of view of the reliability of the software you need. A huge amount of things with driver assist you can write off as 'the driver will take over'. There's not a specific part of you can point to and say 'that's what will stop them', so much as the entire design of the system needs a different level of consideration because it becomes so safety critical, as opposed to safety-enhancing (autonomous driving is probably the most complex safety critical design ever attempted, by at least an order of magnitude. Driver assist is way easier by comparison). It affects the attributes of the sensors, the design of the neural networks interpreting those sensors, the high level design of the decision making of the whole vehicle, the low-level design of the electronics, basically everything. Tesla hasn't put forward a plan which indicates their full autonomous development is going to be anything other than a continuation of their driver assist work (and it's certainly frequently and incorrectly stated that the state of their driver assist puts them ahead in the race to full autonomy), and if they want to be credible to those familiar with the challenges involved, they need to actually show something more than that.

Udik 6y ago
> you can't go smoothly from a driver assist to fully autonomous Ok. So Waymo has developed a 100% reliable full self driving system that handles every possible situation. It was never put in a situation in which it would have given back control to a human driver, because that would have meant "going smoothly" to driver assist mode, and we have already said that this is impossible.
- rcxdude 6y ago
  I was not referring to the dynamic shift while in use, I was referring to the development process. Regardless, the waymo system does not have a driver assist mode, it's either in control or it isn't (and during development it may be overridden by or automatically fall back to a safety driver, who has a job akin to a driving instructor, something which general requires more qualifications and attentiveness than a driver).

jariel 6y ago

The argument is that they are not even in the same category of technology complexity. Which is very reasonable. Tesla's real-world deployment of autopilot is frankly probably an advantage of sorts. And maybe starting with those realities is 'good path'. But it doesn't change the nature of the fact the gap is quite big.

mdorazio 6y ago

A full discussion on this would require an article several times as long as this one as well as some understanding of how vehicle control systems work. I do a fair amount of work with OEMs in this space and to summarize it in a horribly short way: 1) A system that relies primarily on cameras is necessarily limited by camera constraints. No, just because humans primarily rely on vision does not mean an ML model with cameras can do the same and achieve the same results in situations as complex as urban driving. 2) Every major player other than Tesla in this space spends a huge amount of time architecting an entire computing, data exchange, and control platform for autonomous vehicles that is bulky by necessity, but far more robust both in terms of sensor capabilities and ability to cross-check between data inputs to decide what control signals to send. Tesla basically added some cameras and low-resolution radar (they're improving the radar now, but it sure as hell wasn't in the platform they originally promised was FSD-ready) and is hoping it will be enough when everyone else studied the capabilities years ago and decided it wasn't. To summarize, what Tesla has built is a platform with capabilities that very strongly seem to be limited to Level 3 automation, while engineers working on Level 4 long ago concluded that the Tesla approach is insufficient, and then adjusted their own hardware and software strategies accordingly.