“Platooning trucks” A train going down the highway blocking every exit and entrance. That’ll be fun to deal with.
If it does 99% of the rest that you need it to do, then I don’t see a problem. And as I said, that would be when you would be able to take over manually.
The point was was that people seemed to think that the car would just keep going and going until it goes “AHHH I don’t know what to do!”, and dumps the problem on the passenger/driver with no or little notice.
If your SD gets to the end of your drive, and says, “This looks like deep water. I am not programmed to navigate such an obstacle safely. You may either change your navigation to avoid this obstacle, or take over manual control*.”
*during manual control, the software maker will not be liable for the actions of your vehicle. In order to have manual control, you would have to have a special insurance rider that acknowledges this.
Some cars would not have a manual control set up, they wouldn’t need them. Cars that are made for urban areas and the like, where worst case if something is wrong with your car, is that you need to wait a few minutes for another car to come and pick you up.
It would not be an infinite line of trucks, probably 4-5 at most. If that is a problem, then the way that trucks currently drive is a significant problem for you as well.
Sorry k9. But if it did 50% of what I need to do, I’m gonna start to buy lottery tickets. It really is difficult to understand what it’s like to live and drive at 11-2. It’s really not worth the complexity and extra $$ for me. For others, fine. Go for it.
I think the issues that both I and others bring up are important because many engineers and designers are caught in a different bubble than I. Many roads do get closed because of snow. I-70 in particular where I live (I don’t travel that). Getting 30 feet of snow a year makes you appreciate what you need in a car. Being a GIS programmer, I see how long it will take to get any where close to the ability to drive safely, let alone in really bad weather. The head of Google’s autonomous cars agrees with me.
What complexity and extra dollars? You do understand that the business model for these vehicles, due to economic factors, is going to be short and long term leasing. There is no such thing as “owning” a fully autonomous car. Any agreement that allows you to “own” one will have so many caveats it is realistically more like a long term lease. That is, even though some may be “sold”, you will have to constantly spend money on maintenance and checkups by manufacturer authorized mechanics in order for the self driving feature to be kept active. Also, at a particular date, the car company is going to want to deprecate the whole car and scrap it.
Then that sucks too. I don’t lease cars. I rented an apartment once, and then a house. Never again.
There are still adults walking around who voluntarily choose to use flip phones. I was shocked to see an older man in his late 50s who is a high tech worker doing this. Though one issue here is that assuming autonomous cars turn out to be as safe as most experts assume they will be in the long run (10-1000 times safer), this means that every human driver who is choosing to drive is endangering the lives of everyone else.
Only a small amount, but it isn’t zero. Like a surgeon choosing to do a single shot of whiskey every morning before work. They may think they aren’t hurting anyone, and the impairment may be slight, but statistically they are killing more people than if they didn’t.
Yep, people screw up. So do computer programs. I’m really not against this tech. 10-10000 times safer? Ok. Cite please. My Roomba can’t back it’s way out of a corner (I threw it out). I went back to a vacuum that I operate and actually works.
Listen, well, know one has so far, but it’s different where I (and millions of others) live. 26 years of 30 feet of snow a year. Design me a car that can handle that every time I need it and I’ll gladly buy it. Fine if it tells me I have to drive when it can’t figure it out. But then, it’s really like having a 16yo from Florida trying to figure it out. I may as well… just drive.
As a programmer I have to look for every fault. Every little thing that you don’t think a person would do, or want. Oh, they will do it. As soon as you invent an idiot proof program, life invents better idiots. I am the fly in the ointment, the odd circumstance and will continue to be so. It’s sort of my job.
As I have said. Have at it. SDC’s will happen. That’s fine. Just don’t test drive one to my house on August 3rd (our day of summer) and say it’s good to go.
The 10-100 times safer : from Tesla’s own data showing their extremely rudimentary autopilot is about twice as safe, from the nature of a device with 360 degree vision that can compute thousands of courses of actions and gather data on the risks from millions of instances of itself, from programs like AlphaGo Zero, by the same company, which is a path planner (series of Go moves are paths) capable of defeating all humans alive and who will likely every live at that particular problem.
Do you program for avionics or life support equipment?
Do you even understand how modern machine learning systems are architect-ed? There’s very little “if…then” kind of procedural code. There are a series of huge scripts that set up the architecture, at which point all the decisions are made by a pipeline of fuzzy numerical weights and intermediates.
Roomba is a consumer device and the design requirements of that device did not call for high reliability. Also, iRobot did not have the budgets or the personal to develop a reliable system.
So Tesla’s telling us this. Very comforting.
My job in GIS is spatial analysis. Yes, I understand it’s not all Do/While.
The Roomba comment was intended as a joke.
Again, understand that I support the tech. Few seem to understand the hurdles that they have not overcome and perhaps won’t for a long time.
Waymo CEO: Autonomous cars won’t ever be able to drive in all conditions
This is what I’ve been trying to illustrate in this entire thread.
And for fun, to lighten things up a bit.
Our moose. He came by nearly every day in the summer. I named him Frederick.
Ya just never know.
Are you trying to imply that an autonomous car would need special case “moose handling” and not just treat any large lidar return that resembles other large quadrupeds as obstacle.living?
Right, but this thread is whether or not there is demand for SD cars, not whether or not every single person will adopt them. There have been some posters in this thread who have taken that position, but I think it is an odd position to take. Anyone that owns a Ferrari did not buy that car for the practicalities of it, and from my perspective as a somewhat poor person, pretty much anyone who buys anything that’s not an economy car is not doing it for the practical nature of it. Hell, people own planes, helicopters, boats, even submarines. The idea that no one will own their own car just doesn’t fly.
There are also tasks that SD cars are not good for, or defeat the point of. I do enjoy driving, sometimes. I wouldn’t mind taking a nice fast car down to the track, and doing some laps at some ludicrous speed, but a self driving car would defeat most of the enjoyment of it. I know people that like to go “mudding”, off roading in 4x4’s with oversized tires, and they determine how successful their trip was by how long it takes to get unstuck (if how much they talk about how badly they were stuck is any indication), there’s no way an SDC would have any part of that.
But, to the thread, yes, absolutely. There are cost considerations involved, for instance, right now, owning a car is cheaper than taking Uber or Lyft to work every day by a substantial margin. If he margin were smaller, then I would probably switch, even if it were a bit more expensive. Not having the stress of having to maintain my car, not to mention the stress of driving along a bunch of idiots who care more about their text message than about the safety of themselves and others, would be worth a bit of of premium. If they came down to being less than the cost of car ownership, then I am on that.
90% of the time, my car sits idle, in my garage or in the parking lot at work. 95% of the time it is in use, it is covering the same exact 10 mile stretch of clear pavement in nominal weather conditions. The only concern I have is when the weather is poor, and I honestly think that it is the human drivers that are going to have more problems with slick surfaces and low visibility and lightly flooded roadways than an SDC would.
When conditions are much more unfavorable, when there is snow piled up on the road to dangerous depths, or when the road is flooded or washed out, the SDC will probably not be able to handle it, but it will probably still handle it better than 90% of drivers, in that it will say “No”, as opposed to “Let’s take a risk.”
The idea that AV’s cannot handle all conditions isn’t really a non-starter for me, as most drivers cannot handle all conditions.
No, I was just sharing a picture of a moose in our yard. I think it’s kinda cute.
Or you just whooosed me.
(Photo Bucket sent me yet another email about how I should pay for their service, and I decided to browse my pictures)
You where the poster that said that the best way to get out of a stuck snow condition is to have the tires spin. Please correct me if I’m wrong.
Perhaps a SDC would be good for you. Make no mistake, I see plenty of people in rental cars that really, really are in over their head.
That’s fine. As I’ve said, go for it. I am not against this tech unless it is forced upon me.
I am not against SDC’s. It would work great for many. For others, they are gonna suck.
My life and my job revolves around … um, well that’s not gonna work. And so here I am pointing out flaws.
I was being very simplistic, but yes, the traction control prevents the tires from spinning at all. If it detects any slippage, it stops the tires. I’ve seen people just spin their tires, and think that if they can just spin them fast enough, then they will get unstuck, (I did try that once at 17) and that is not what I am talking about. I am saying that the traction control, when stuck in snow, will prevent your tires from moving at all, as any movement will produce slippage, and get shut down.
That’s why there are handy functions to turn it off (though on my current car, it’s under two sub menus, with other cars, I could assign it to a button on the steering wheel.) When it is snowy, and the roads are slick, I like to turn it off when I am stopped at intersections, or else it won’t want to go anywhere when I step on the gas. But I turn it back on when I get moving, as it will keep my car moving in a straight line. (Especially on the rear wheel drive Caddy, front wheel, traction control almost never comes on.)
Like I said, depends on price point. It is doubtful that I would be able to afford to own one of my own any time soon, as we are probably talking about fairly high end luxury vehicle prices, but the lack of labor should bring the price of rides down enough that it makes sense. Maybe I keep my current car and take it out on days when I feel like driving, maybe I don’t. Maybe when my current car goes kaput, I get a new one, maybe I won’t.
Far too many variables to say for sure.
There is a good chance that they would be forced on you in certain situations. New York City doesn’t allow people under 18 to drive in the city. There are other cities with similar type restrictions. Cities would be the first places where SDC’s make the most sense, from a technical sense of being the easiest to navigate, and having the highest density of potential customers, and from the need to lessen and lighten traffic and accidents. It would not be long after the implementation of a reliable SDC that many cities may not allow you to drive manually within their limits, or on certain roads.
If you go into a city, it is quite likely that you will either have to have automatic controls on your car, or switch to one that does. I see interstates going that route fairly shortly after, with dense suburban surface streets probably being the last. Low density suburban and rural areas will probably allow manual driving for a long time, or even indefinitely.
And they do have flaws, some of which should be easy to fix, some of which may be harder, some of which may be challenging, and some of which may never be resolved.
I still think that current state of the art is probably better than 60% of the drivers out there, if only because it doesn’t get distracted, drunk, angry, tired or stupid. I don’t think it will be hard to get it to be able to handle 95% of conditions taht are found on paved roads, and do so better than 95% of drivers.
Some other special tasks may take longer, or not ever get covered as they are so rarely encountered (by people in general, not individuals who live near these conditions) that they never really get properly addressed.
…the thread is actually about “is there any market research that indicates serious DEMAND for self driving cars?” There is obviously demand. But can you quantify that demand?
Uh huh. It is complicated and I understand hard to explain. But much of what you wrote is just plain wrong.
I’m really not trying to pick on you or others, but the lack of knowledge of what real snow driving here is a bit stunning.
[sidenote]k9, your business is working with dogs. That’s wonderful. My Wife and I have rescued 4 dogs now. It’s one of the many reasons we drive SUV’s[/sn]
We are getting away from the intent of this thread. I only offer data and information that a poll may not ask. For myself, a SDC is way, way far away.
Personally? No.
I could google about to see if there are any good polls out about it.
Or you could check out the ones I posted upthread. 
And if I understand correctly, Waymo is considered the best of the outfits trying to develop autonomous autos, perhaps even head and shoulders ahead of the pack. If this is what Waymo is saying, autonomous vehicles are probably a long, long way from arriving at the dealership.