My concern over Waymo's solution is they rely on remote operators to take over if the car gets confused.
But, how long is the latency?
Now, this isn't to say I disagree with it. There really is no other way to get autonomous cars on the road in the short term than to do that. It's also extremely important from a society standpoint to get people used to autonomous cars.
But, I think it's a stop gap solution.
I don't believe that any truly autonomous car is possible without car to car communication, and intelligent roads.
If you don't have those two you've thrown away one of the biggest benefits of automation. That is to make the road part of the entire system. It also makes it way safer since the car doesn't have to see over a crest, or around the massive truck in front of it.
There also construction zones that can be easily programmed to tell the car exactly how to handle it. Heck they should do this for human drivers as well because sometimes they completely fail to communicate to human drivers what's going on. As an example last night I was driving north on I5, and I saw a flashing sign that said the 3 left lanes were closed ahead for construction. So I moved myself over, and everything was fine. Only a few miles later I entered into another construction zone and the sign said the left lane was closed ahead. So again I got myself over. But, then I noticed up just ahead the cones seemed to be going into my lane as if my lane was closed. So wtf? It didn't said 2 left lanes, Luckily there wasn't anyone next to me so I got over.
Intelligent roads is also a matter of efficiency. What's the point of autonomous driving if we can't take advantage of it? Where the cars can communicate to each other, and the stop lights can be intelligent. Where they communicate with other stop lights to improve efficiency.
I also find myself a bit longing every time I look at the Model S instrument cluster. I do because it shows me the lines which would be extremely useful in bad weather, glare, road construction when the lines are bad, etc. But, it doesn't. It just goes blank as if somehow as a human being I'm supposed to see the lines?
Humans do have a great ability to understand how other humans drive, and to react to situations before the unfold. But, humans also tend to make believe a lot. Where we go "oh, crap I can't see anything so I'm just going to slow down a bit and pray"..
This is a great post because people actually think that in 3 months All Tesla will be Level 5 and others think by June 2018 they will all be level 5. Its the most delusional thing i have ever heard. Almost as good as the people who thought that 150,000 model 3 will be made in 2017 cause elon said so (while i warned them that Tesla would be lucky to even make 10k model 3s and was laughed at) and that the model 3 will self drive to their house after the AP2 announcement.
As someone said: "One of the most amazing things about Tesla is how Tesla fans manage to hold thoughts which are so patently false, against all types of evidence."
Tesla is literally a breed ground for myth after myth.
Not only is Waymo going live this fall, but you have people who actually think that a 3,000 mile highway demo with 10 miles of surface street is supposed to prove Tesla is the leader.
When in reality, if you are still doing highway demos now you are 5 years late! Get Rekt! You Already Lost!
Such facts such as the magnitude of highway driving vs suburb driving vs urban driving literally escape Tesla fans. Even the service Waymo is running in phoenix, though unimaginably more complicated than highway driving is nothing compared to urban driving in a city like SF.
Tesla fans will hate this, they will call me anti-Tesla but I'm only speaking the truth.
If you think that Tesla (and i know many of you by name who do) will have Level 5 cars that can do this by June 2018 then you are delusional.
While Tesla is struggling to match level 2 lane keeping and adaptive cruise control and panicking at highway autonomy. Guys like GM Cruise and tackling the difficult part of self driving. Actually self driving on the streets.
And no you don't win based on participation, this isn't middle school. You people think just because Tesla mounted a Nvidia card on their car and strapped acouple cameras they are now world leaders in SDC software? laughable.
This is what driving on the street looks like not the empty roads of interstate highways. (The predicted future path and multiple possible interactions of every person, car, and cyclist are calculated 10 times per second.)
Navigating a 6 way intersection with damaged non-functioning traffic lights
Over taking big trucks.
Handling Human controlled construction zones
These are frequent occurrences in any densed urban city (per 1,000 miles)
Why testing self-driving cars in SF is challenging but necessary
Tesla leads in a sense that its final hardware design is done and sold to the public while others are still having all those sensors poking out. Some still have big, gigantic computers in the trunk.
Tesla leads in a sense that AP2 improves very quickly since its first release to first 1,000 owners in 12/31/2016.
It is true that other prototypes are much better current Tesla AP2. Waymo has been a leader in autonomous technology but like others, they are not released for public use.
It took AP1 one year to have its autosteer released so I expect AP2 will surpass it in 1 year.
Waymo's high definition maps are done by sending its own employees to do the work. Tesla leads in this kind of data collection job because it doesn't depend on a small numbers of employees to do the mapping but its entire owners' fleet since October 2014.
They don't say Waymo leads because the only lead they have is on prototypes. There are probably hundreds of self driving prototypes out there from various companies (even from more than a decade ago), but largely irrelevant if the timeline for public release is indefinite. The closest thing Waymo have to public release are the slow moving neighborhood vehicles (they are still controlled trials though), but those are much more limited in practical use (only work in certain geographic areas, very slow).
Tesla is the only one with a system in widespread consumer hands (although software only L2 at the moment). Maybe someone else will take the lead (for example GM's supercruise), but so far Tesla is still in the front.
You two can't actually believe these posts anymore do your..not in the face of mounting evidence?
