I read somewhere that one of these companies (I think Uber) was thinking of a model for self-driving long-distance semi trucks in which the truck would be autonomous on the highways, but when it got into town, a human operator would take over, steering the car from a remote location. That sort of model seems like it’s more likely to happen soon.
Dude, it happens all the time at car show press events. It happened to us when GM introduced the Volt at the Detroit auto show back in 2007 or 2008. Four of us had to push it out onto the stage and we just left it there after the event until they cleared everyone out of the booth. It was hilarious, but hardly an indication that a company is “struggling”.
GM needed its $13B bailout in 2008, so it doesn’t exactly qualify as a “non-struggling” company at that time.
Sort of like a Predator Drone. This is not necessarily comforting.
Now that was a bit of sarcasm. But not entirely. I’m pretty cool with automation. But when it takes three dam days to upgrade a piece of software built by the biggest GIS software company in the world, well, it gives me pause to put that behind a 40 ton truck.
“Yes, we forgot to put that step in the instructions”
Thank god it was a development machine. Every time I have to update software it gives me chills.
All you naysayers are hilarious.
Y’all may commence again with claiming that self-driving cars “are still decades away” and in trying to explain how these self-driving cars are not, in fact, self-driving cars.
They will work in some limited circumstances soon. But a true self driving car is at least, at least 10 years away.
Bo, no one disputes that there are cars today that can drive autonomously under certain conditions. But you’ve been told several times that that is not what people in this thread are discussing. “Decades” refers to AVs available to the average consumer that can legally go on any road in any conditions without a backup driver.
If you want to use all your little anecdotes to argue why the timeframe for that is shorter, go for it. But these constant assertions that AVs are already here is a strawman based on a definition only you are using. If that’s the only definition in your mind, then your work here is done - we already know what’s here today. The discussion has moved past that.
A few test cars in Arizona or certain other cities is far, far short of autonomous cars that can drive themselves in rural PA, MI or NY on roads without white lines, no shoulders, in the dark. Add snow, and even highways would be difficult. And then you have temporary lane closings, stalled cars and detours. I think two or three decades is optimistic for even modest use.
But are the primary difficulties now technological or legal?
There are plenty of both, and I’m not sure you can say one is primary over the other.
And I’d throw in a third category, if I could think of the right name for it. A major challenge is AVs that can work alongside human drivers, who are unpredictable and will take advantage of cautious AVs. Figuring out how to program in the rules is simple compared to figuring out what those rules should be. What’s the term - social? Behavioral?
In a magical world where we could switch to 100% AVs with no humans on the road, AVs would probably be fully viable in a couple of years.
Did you just say that once a driver’s ed student passes his test, he requires no further supervision? Legally, you are correct, as far as supervision goes, but a 16 year old driver still has many restrictions. Once you have a full license, you no longer legally need a supervisor, but do you really think that that newly graduated driver’s ed student is going to be as good a driver as someone who has been driving for a decade or two?
Self driving cars are like taking that driver’s ed student, who has already passed the test, and continuing to supervise him. Every mistake he makes, you tell him what he did wrong, and he remembers it. Any accident or incident or anomaly is also learned from.
And the great thing is, is that it is not millions of driver’s ed students, each making a mistake and hopefully learning from it independently, but just one, who gets thousands of hours of experience in just a day, and is able to share it with all others.
Your “poor depth perception” comment makes no sense, as a lidar system, or really any computer controlled ranging system, is going to be far more accurate than the eyeball mark 1.
As far as the “take over” bit, that’s just during the learning curve. As a passenger, are you expected to take over if the driver of the car makes a mistake, does not see an obstruction, or has a sudden medical crisis?
The only times you would need to take over would be well known in advance, as you would be aware that you are going off-roading, or even bad gravel roads, before you left the main road. The chances that, should you be paying close attention to the road, that you would see something missed by the car’s sensors are very remote.
Right, which is why during this testing and learning phase, they should be driven by driver’s ed teachers or the equivalent. Do you think that this will be the state of the technology forever, or can you see that this is just a testing and training phase?
As long as a person is needed to be i the loop, you are correct, people won’t much like them, and I don’t think that they should be sold as self driving. Once people are out of the loop, and are only needed for when you are driving up your snow covered driveway that you don’t think the autopilot can handle, it’ll be a bit more accepted.
Google self-driving spinoff Waymo to put factory in Michigan.
I look forward to what I am sure will be many, many replies detailing why this is not a real auto plant, how the plant is not producing self-driving cars immediately and how even if it were, they wouldn’t be real self-driving cars.
A $13.6 million plant with 100 employees that they hope to open this year at a location they’ve yet to determine. Does this even need a reply?
Apparently it does, and as I predicted, it was an attempt to denigrate and belittle the story. And hey, I’m not disappointed: the very first reply denies that it is even a plan.
What do you mean, as you predicted? I was responding specifically to your well-poisoning antagonistic prediction of the “many, many replies” you expected. Clever trap, though.
Bo, see my post #267. Your post is unnecessarily antagonistic and ignores what people keep explaining to you.
Okay, I’ll be one of those. You can’t build a car factory for $13.6 million dollars, unless you redefine what it is to build an autonomous vehicle. If you buy everything pre-made (like they currently do) and then install their own sensors and hardware, then maybe. But then do you consider companies that take chassis cabs and add a box to make a box truck as a vehicle manufacturer?
If you read the article, it’s not a car assembly plant nor does anyone claim that it is. It’s a conversion facility that installs self-driving components on Chrysler Pacifica minivans and Jaguar electric SUVs.
What I came here to say is that for those claiming that regulatory obstacles will delay self-driving cars, now that it’s 2019 it’s legal to test autonomous self-driving cars on public roads in Ontario with no driver behind the wheel – and this is a province with notoriously conservative driver laws.
One such vehicle will be the University of Waterloo’s “autonomoose”, which has mastered most of the basics and can even drive in snow and various forms of bad weather, though not severe weather. Even now, the challenges with the UW car are all about exceptions, like broken traffic lights or the anomalies of construction zones. Even if it has 95% of the capabilities it needs, it’s going to take a lot of effort to span that last 5%. And a lot of fine-tuning, as today these cars tend to be very meek, yielding (as one researcher put it) to anything and everything, which makes it hard for them to do things like changing lanes in heavy traffic with aggressive drivers. OTOH, trust me, not even a human driver can do this out of Logan Airport approaching one of the tunnels heading into Boston. Boston drivers are insane. 
It reminds me of Trump claiming that he saved 1,100 jobs at the Indiana air conditioning factory, and the die-hards just ate it all up.
Then a few months later, when the ruckus died down, jobs were sent to Mexico anyways.
When this auto plant starts putting out modified cars, then let’s talk. That date could be in two months, or 2026. Who the hell knows?
Actually, if you read the article like I did, it doesn’t say that’s what the plant is; it says that that’s what they currently do. The nature of the plant isn’t mentioned.