Imagine how much more GDP the US would have had if they had done that...
edit: Of course, Europeans also use that extra time to enjoy life versus driving GDP that due to increasing wealth inequality won't benefit them in any tangible way.
There are 250M cars on the road in Europe. Clogging up public space and requiring human labor to be driven around. There are cars everywhere in Germany.
A self-driving car can take you exactly from where you are to where you want to go. Right now.
Public transport requires you to walk to an entry point, wait and then walk to your destination.
Self-driving cars will make things way more efficient. Similar how TCP/IP packets are more efficient than TV broadcasts, which required you to plan when to watch what and then wait for it.
So if you propose that Europe should try to achieve efficiency via more public transport, I think that is a losing battle. Electric self-driving cars are the future.
Road capacity has a physical limit. Unlike data you cannot convert cars into light of various wavelengths to pack more into the same physical space. Well you can but that's called public transit. In other words, public transportation has a higher capacity per road/rail area than cars ever will. We don't live in the metaverse after all.
As a result more people using cars (self-driving or not) means you hit the road capacity much more often. That means more congestion and longer trips. Any efficiency you gain by going point-to-point is utterly lost in the increased congestion and traffic.
As the other person noted, European cities are much denser and have significantly less road area than US cities. Thus more cars means significantly less efficiency as you spend more time in traffic.
- A train will take you right into the city center, sometimes even where cars are not allowed. OTOH entering the city from suburbia can take an hour of being stuck in traffic.
- metro and trams can skip traffic. Bus lanes are common. There are no "stroads". Average car speed in London is 12mph. That's slower than a bike.
- Finding a parking place can easily take longer than waiting for public transport, and it will probably be an underground garage, far away from your destination. Bus stops are much denser than parking garages, so they usually involve less walking.
The US experience of driving on an 8-lane highway and arriving at a mandatory-parking-quota asphalt desert in front of your destination is rare in Europe. The closest we get is a 3-lane motorway to an IKEA on the city outskirts.
And without parking cars, streets become wider. Single lane becomes double lane.
In some dense European cities there's physically not enough road surface to put every traveller in a car, self-driving or not.
https://en.m.wikipedia.org/wiki/Route_capacity#/media/File%3...
Public transportation increases the potential commuter density per area of road/rail while self-driving doesn't.
But if a bus, a tram or a train would have a self driving... but that smells like communism to our techbros from Silicon Valley.
Musk said in 2 years, in 2012. Two more weeks brothers, two more weeks ... lol
https://www.caranddriver.com/reviews/a39966189/2022-mercedes...
All I have seen Mercedes do could have been accomplished by manual programming. But manual programming does not scale. Look at the progress Tesla made from FSD 11 to FSD 12. Nothing indicates that Mercedes is capable of this type of progress.
That and the fact that Elon has spent the last decade saying they're 6-12 months from actual full self driving.
Tesla - 0 disengagements.
MB - 50 disengagements.
This is not even fair comparison. MB can take all the responsibility they want when no one is going to use their system.
LOOK AT THE PROGRESS!!!1
And with FSD 34 it will be able to take the easiest highway exit ramp there is on this planet, Mountain View on HWY101, without killing you.
And what does "manual programming" even mean? Do you REALLY believe the xbox360-grade hardware in the tesla is doing AI inferencing? It's not. It's a heuristic.
The biggest problem is really how do you get to 10^n miles per disengagement, for n>=5. Waymo is kinda getting there, Tesla isn't anywhere near that today.
Getting there is really hard, because that's when you get all of the long tail events like bears, moose, wild turkeys, horse mounted police officers, costume conventions, pickup trucks carrying traffic cones and road signs, flooded streets, construction pilot cars, vehicles driving the wrong way on the highway, downed electric poles, NYC steam plumes, and tons of other scenarios. Highway driving in nice and sunny conditions is easy compared to that.
Do you know of any reverse engineering that proves that there really is running anything in regards of inferencing on the NPUs?
Also, just as you said - there are tons of corner cases in the real world, especially once you aren't on a 10-lane US highway which has been designed for monster trucks driven by 16 year olds (no offence) but one of the roundabouts of hell in Paris.
Where would the training data been coming from?
So, I have my doubts.
During summer, there is a red flower growing near the entrance of my parking garage. It constantly is seen as a red light, and the entrance of my garage is often mistaken for a huge truck suddenly magically appearing. Again: Nobody would use a Captcha these days: "Is this a red flower or a traffic light?".
Again, smells like heuristic. "Amount of red pixels in a certain form and spot".
Typically, inference in a machine learning context means feeding a model some input and looking at its output. I'm pretty sure that they are running some model on the vehicle that takes pixels as input and says this part of the image is a car/truck/traffic sign/lane line/etc. It might be misclassifying things (eg. the flower as a red light), but would still be running some kind of model.
As you point out though, the model only seems to do some simple object detection, but doesn't have much of an understanding of what it sees (eg. does it make sense that there would be a traffic light at this location). There are plenty of videos of it getting confused by all kinds of situations (eg this one from a few years ago https://www.businessinsider.com/tesla-fsd-full-self-driving-... ).
Firstly the road structure is very different to the US, streets are much narrower and less car-centric so models are harder to train; secondly regulating it is much harder because the EU is a collection of countries, which is very different to a collection of states.
Thirdly non-car alternatives like trains and ebikes are a lot more viable due to a higher population density, many big cities (such as Paris) are pushing for better cycling infrastructure rather than better layouts for cars.
I believe VW still has a group working on this.
https://www.tesla.com/de_DE/autopilot
But this feature simply does not exist. It's a myth spread by Tesla fan boys who do not want to admit they got screwed over.
To be clear: If I drive a very easy street through my town, which is near a school, my Tesla - just like EVERY Tesla - is able to detect about 90% of trash bins on the side of the road (impressive detection rate), but only about 60% of pupils crossing the road.
If I would use Tesla's "full self driving" feature, I would be on average killing about 20 school children EVERY SINGLE MORNING.
So, no, it's not that.
There is absolutely no actual data to verify this claim. How would you know how many "billions" of miles have been driven with "FSD"?
Tesla isn't publishing any data. They just have been caught actually hiding crash report data to the NHTSA.
"My" anecdotical data is simply watching the amount of trash bins vs kids vs other objects the Tesla I am driving is able to detect. But that is still better linking to an unrelated Forbes article talking about damage done to a car when crashing it into a wall, which does not have anything to do with "FSD" in any car.
Actual facts here instead of "Elon says":
https://static.nhtsa.gov/odi/inv/2022/INCR-EA22002-14496.pdf
I beg you pardon?
https://etsc.eu/europes-first-cars-with-level-3-automated-dr....
/thread
Mercedes' "level 3" thing has been around for a couple of years in, at least, Germany. I'm not sure that there's all that much to "fall behind on", though.
> If we value an hour at $20, that is $1.8T per year which can be automated.
Can it, tho? Like, the last 10 years of promises about self-driving cars would indicate no, at least not any time soon.
And they don't offer that in Europe.
As far as I know, there is no self-driving in Europe at all. No Tesla, no Waymo, no nothing.
Will Europe fall behind with self-driving like it did with the internet?
There are a similar number of cars on the road in Europe and the US. About 250 million. I wonder how much they are driven? Maybe one hour per day? That would be 250M*365 = 91B man hours per year. If we value an hour at $20, that is $1.8T per year which can be automated. That is about 10% of the European GDP. Plus many other benefits of self-driving cars, like more public space because fewer cars are needed, and they can park outside the center.