I don't want GM to ever be a "Tesla alternative ". Musk was called out By Robin Zeng who is the head of CATL who makes all of the batteries for the Chinese Teslas. He asked why Musk continues to overpromise and under deliver? Musks classic stupid response "I always say two years out more than that is infinity." We don't need any more vehicle manufacturers electing ceos that are all over the map.
$10 Billion Down The Drain - The Cruise Is Sunk
- Thread starter macarose
- Start date
GM has been working in “Ultra Cruise”, the successor to Super Cruise for a while. I’m sure they will just focus on that with this tech. Ultra Cruise was renamed to just stick with “Super Cruise”.
https://www.theverge.com/2024/1/22/24046928/gm-ultra-cruise-merge-super-cruise-adas
https://www.theverge.com/2024/1/22/24046928/gm-ultra-cruise-merge-super-cruise-adas
Ive spent more hours than i care to recount listening to researchers on this lately, for the most part they don't agree at all.
I ( and the researchers) agree that a dedicated lane or a "smart highway" would make it more viable. It is not happening and even if we started tomorrow it would take years into decades to build out the needed infrastructure.
I agree there has to be support, i don't think anyone will get there in any amount of time that could be termed "soon", it think europe and japan embrace rail much better and probably the cost of car owner ship helps that.
Another interesting factoid brought up by VTTI, apparently as experience with ADAS rises the effectiveness is eroded by drivers shifting attention to other task and recovery time increases.
One of these researchers also went on a thinly disguised anti a certain very rich person rant about how this persons place in the new regime was going to create pandemonium with a certain brand of car roaming around all by itself with no consequences... Irreally don't care to say more and probably cant, but i would not necessarily assume support for this sort of thing will be less than current.
Also note when i say it is a ways down the road, i mean a situation where you can get in a self driving car and it drives you from ny to grandmas house in OP AL, it may very well be the case one will take you across nY city, or London or Paris or something where it is tightly controlled and regimented.
If you watch some stuff on the tube of you, apparently even where these or waymo were/are deployed it is pretty common for them to do really stupid stuff...
In my mind, something cheaper and simple like a wire down the centre of each lane, or maybe plop some rfid chips every 20 yards to identify each lane, so each car knows where it is at all times, then add some communications between cars, so they all know where every other car is within a mile or whatever makes sense. Most cars will have a planned destination, and planned route to get there, then run it like a train system, so they aren't trying to interpret visually or with lidar to figure out what all the other cars are doing. Trying to simulate a human brain and eyes is not cheap or possible, so far.
I don't think the computing for a huge train system would be that difficult? Maybe it has to be distributed between vehicles, but computing power is cheap when it doesn't have to be tiny or all that efficient..
Making the system hacker proof is probably the biggest challenge?
Even old "dumb" cars would need the position tracking and communication system installed, to let every other automatic car know what the car with the crazy human is doing.
Maybe some of our resident programmers could give some insight in the challenges in programming the operating system for something like this?
I don't disagree that something like that could work, there's still a lot of questions around it the foremost i imagine being "who is paying".
I don't pretend to be an expert either, but certainly at least from what ive heard there are big gaps in research and not much support for an infrastructure based system. ( I actually think that is where a lot of people would like to see it go)
Based on most of what ive heard i don't think anyone really thinks level 5 is attainable soon, and i'm not sure we wouldn't all be better off to just go back to like level 1.5.
I forget what they called it but apparently there is a concept that there is a certain level of risk considered acceptable and if driver aids reduce the risk too far, the brain will find other things (like read a book) to bring the risk back up. This sounds a little crazy but there's apparently research to back it up, also as driving aids use and familiarity go up, the time of off the road glances increases quite markedly.
Folks who are really interested in this stuff should look into places like:
https://mcity.umich.edu/what-we-do/mcity-test-facility/
and
https://www.vtti.vt.edu/
those should get you started toward finding some of the research.
That's not what it looks like? -resident of the Great Plains
Cruise was never going to be viable commercially for anything other than in city taxi service where the enormous external sensors didn't effect highway mileage, and you could operate in a fully pre mapped geofenced environment.
The Cruise was GM’s version of a “ World Car”. They thought they could cut costs and manufacture and sell the car in many countries. Surprisingly, they sold just as many in China as in the USA. Here’s a couple of screen shots.
Different Cruise.
Oh yeah!
Well, it’s sunk too.
4WD
$50 site donor 2025
That is an improvement - now if some horseradish floats by
I agree with you. People are so accustomed to a quick buck these days that they forget this is a typical part of every industry and see it as a flaw, when it’s just part of business. R&D usually pays for itself long term which isn’t as sexy as 100x profits in your pocket next year.
They'll repeat history by expecting another federal gov't bailout... GM= government motors.
