If that is the only use case Mercedes feel their safety assurance can justify what does that say about other self driving cars for which their manufacturers will not accept liability?
It says that they’re a responsible company that isn’t comfortable playing fast and loose with your safety. Unlike some US companies that don’t come from a background of safety but a background of “move fast and break things” and a business model of regulatory arbitrage.
But this is quite a bad product in that respect. If this is all it takes, Tesla would implement the same restrictions and get the same fanfare, but alas they're not constantly aiming for "PR stunt" levels of driving.
You're saying that Tesla of all companies isn't aiming for PR stunts with their automated driving? Tesla, the company that calls their level 2 automation "Full Self Driving" isn't out for PR stunts? I would be hard-pressed to name many companies that seem more prone to PR stunts than Tesla.
Assuming that Mercedes Benz (and all other manufacturers at that) know their shit, which they do, it says more about the self-driving competition and self-driving tech in general thann it does about Mercedes' offering.
> I personally would rather have a system that I can actually use when I want even if that means I need to accept liability while using it.
It's all a trade-off. My accident-rate thus far is one serious accident in over 700000km of driving. My fender-bender rate is three FBs in over 700000km of driving.
My understanding of the Tesla system[1] is that it requires roughly one intervention every ~5000km of driving in order to avoid an accident. For me this is an unacceptably high risk, because not intervening in 4999km will definitely (100% certainty) mean that I will be in a poor position to react when the intervention is necessary.
Now, you might claim that the driver has to be alert while not in control for 4999km to avoid the accident on the 5000km mark, but if drivers were that good at being alert while not engaged with the act of driving, then the self-driving system is redundant anyway.
[1] I read the stats a long time ago, so maybe they've changed.
From my point of view, the actual act of driving is not that difficult (after the first 50k miles or so). The issue is the mental effort required to continually pay attention to what's going on to drive safely.
Tesla's system still requires me to pay attention to what's happening to exactly the same degree as normal because I might need to intervene (and in fact makes it harder to be ready to do so). All it does is take away the (to me) trivial aspects of pressing a pedal and turning a steering wheel.
Whereas this system introduces a set of circumstances in which I don't need to drive at all. No need to pay attention at all. And it's the most tedious form of driving there is - stop start traffic on a motorway.
> And it's the most tedious form of driving there is - stop start traffic on a motorway.
As someone who has used autopilot in the circumstances you describe, I'd be shocked if you need to pay more attention during autopilot than the Mercedes system. Autopilot generally requires intervention in danger situations, which are incredibly rare in stop/start traffic.
I'm not sure that I understand your point. Which party takes responsibility for an accident says nothing about the frequency of accidents or the kind of threat this car would pose to other drivers.
Of course it does (indirectly). Tesla for example can't take responsibility, since they know that their "full self drive" or "autopilot" systems are never reliable, in any driving circumstance.
What you're saying is like "the engineers that built this bridge never drive over it, but that doesn't mean it's shoddy" - technically correct, but almost certainly wrong in practice.
You are assuming the motivations behind these decisions are purely based on safety rather than a philosphical difference in approaches.
From a practical standpoint, liability needs to be all or nothing. You can't have a driver worrying about whether they are going 40mph or 41mph. You can't immedaitely give the driver liability if it starts to drizzle or the sun sets.
Mercedes is taking the approach that they are always responsibile. Other manufacturers are taking the approach that the driver is always responsible. The end result is that the Mercedes system is much more conservative in how it can be used. This says nothing about the quality of their technology in comparison to their competitors. It simply says they are focusing on the easiest problems first while their competitors are taking a more holistic approach trying to design a system that has more broad usability.
As everyone knows by now, literal full self driving (as in get in your car, tell it to take you to the other end of the country and wake you up when it gets there) is entirely out of reach to current technology, and will stay out of reach until we design new sensors and possibly general AI*.
So, the current goals must be to achieve something similar in certain well defined limited conditions, and with reliable automatic checking that you are still within those conditions - hopefully conditions that one is actually likely to encounter. Until we have that, letting self driving cars on public roads is a menace.
Current self driving cars are at best at the level of a driver going through their first driving lessons, and one with very bad eyesight at that. Having a human act as the driving instructor, theoretically prepared to step in whenever the AI makes a silly mistake, is not enough to make these cars as safe as the average (non-drunk, non-sleep-deprived) human driver.
What Mercedes seems to be doing is responsibly pushing the state of the art further. Having a car that is safer than a human driver without depending on your constant vigillance is a huge step forward. Obviously, this only works in certain conditions, but the car itself detects when those conditions are no longer met, and gives you quite ample warning to assume back control.
* Elon's shameless lies about having your Tesla act as a self driving taxi and generate a profit for you while you work in the coming years have well been put to rest.
The point is that if the designers are not even confident in saying "this works without a hitch under X and Y circumstances", allowing its use on public roads at all (and you choosing to use it) are bad ideas.
Yes, but you are missing the essential point of this move.
They are saying that they have enough confidence in the system that they will pay for anything that goes wrong. This means that they have run the calcs to work out how much its going to cost.
Tesla have basically gone: "fuck me its bad, lets just legal boilerplate ourselves out of the consequences. Oh and charge people to QA our shit"
