Well there are doubts we'll have much of a collector market after 2030 anyway. Insurance models have cars that aren't self driving off the road by then as too expensive to insure. But, assuming Tesla survives, at least their cars can be upgraded so should be able to stay on the road. Inability to upgrade traditional cars is going to be a problem. There have been several attempts to create a retrofit but the lack of common interface standards at a deep level plus the lack of modularity in the internal systems and the huge focus on car companies wanting us to replace not update has been problematic.

 

It's a good thing we have thousands of years to worry about that.

 

Fixed, thanks. Teach me to type and not proof read. But on the collectors cars, if we look back the XKE, I owned 2, is really the latest Jaguar that has come up in value nicely. But had I kept my '67 which I bought for $2,500 I'd have to put $50K to $60K into it to have a $80K+ car. I'd have made $20 to $30 but had I done the same with an XJS, the next model I'd likely be down the same amount and it is doubtful that car will ever close on the price of the E, 120s are pricey and a decade older but more expensive to restore as well I'm told. If you go newer XK8s still haven't bottomed yet and some of those are getting close to 20 years old and given there numbers it seems doubtful they'll ever be worth more than it would cost to restore one. You can make money flipping classics but keeping a car long enough to do it is iffy, and think about it, it really took nearly 40 years for the E to really start to move and while I may be alive in 40 years I doubt I'll be keeping my F that long even if self-driving cars don't knock it out. I was going to do a comparison on used car prices on eBay but prices on the Teslas are all over the map and there is a massive jump between the 2014 and 2015, something like $30K which may have to do with the motor upgrades but I'd have to do a model by model check and don't have time between planes. Good news is it looks like the prices on the V8S have stabilized, they seem to be pretty close to what they were last year if I recall correctly. Makes those a decent buy, I didn't look at the others. Anyway something to noodle on.

 

Do you have a source for 2030? Given that we are sitting in 2017 and there are no production self-driving cars I think 2030 as an obsolescence date for human driven cars is very optimistic.

Will the tech be good enough by then? Very very likely. Will much of the current driving population be able to afford a newish car? Very unlikely.

 

It is based on a number of factors,I'm one of the analysts published on this). Uber being almost exclusively self-driving by 2025 (so there is an alternative), insurance models that have costs for self-driving only cars dropping by 90% and "guardian angel" cars dropping by 60% or better. Similar models showcasing that, conversely, human driven car insurance will increase from 100% to 10x (really depends on if human cars are allowed on freeways by then), and historical buying behavior which suggests those with older cars can't afford a large increase in insurance costs (so will switch to uber). We are updating for recent millennial behavior which suggests the Uber trend was too conservative (kids in cities are choosing to not get driver's licenses at an increasing rate). But we'll have a far better idea of how realistic this is when we have a large number of these cars on the road in 2019.

It is interesting to note that the two models are currently in contention. The true self-driving the cars will lose any ability for a driver because testing has shows that the driver is more likely to cause an accident than correct for one. Initially it was thought that, if the car ran into trouble, it would alert and given driving control back to the human but then the engineers realized that if the computer couldn't deal with the problem the driver (who might be reading, sleeping, or watching a movie) couldn't move fast enough to assess the situation and respond so redundancy is being built up in the systems. But talk about your "oh crap moment". Toyota is the firm that brought up Guardian Angel, but it is kind of like a car seat for your kid. All drive by wire, the computer is always ready and if you do something stupid it will simply prevent it. But, like traction control, you could turn it off but, if you do, it could invalidate your insurance. The tech is pretty well cooked the insurance and legal portions not so much. I expect, as a result, the "turn it off" part will likely either not make it to market or be short lived. We'll see.

 

What happens to the majority of the country (by area) in which Uber/Lyft do not have a significant presence?

 

Dont worry, he neglected to mention the fact that autonomous cars have no idea what to do when it snows also. We're much further away from level 5 autonomy at a global perspective. It will be decades before cars don't come with steering wheels.

 

Most of us are focused on city centers, snow, so far, hasn't been that big an issue in testing. It is ice that may be the bigger problem though the cars, when they sense ice is likely, slow way down and the computer's ability to respond to ice is pretty good (they don't make the typical human mistakes) however black ice on a hill where the car goes ballistic is a problem regardless of what is driving it. Agreed, much like horses hung in there in the country far longer than in cities and freeways, we may see areas that move more slowly. Insurance may still be the big driver though as those vastly higher rates will certainly change behavior. But agree, there are some hills and valleys here we can't yet lock down. But, as far as weather, watch how the big truck trials go, given those things are expected to go nearly 24/7 and the liability is far higher, the weather testing looks to be far more rigorous. The big advantage really is that the computers just don't do stupid stuff like drive beyond the vehicle's ability to see or react. It shouldn't be too surprising that this behavior seems to reduce accidents in all kinds of conditions dramatically.

 

There are too many questions left unsolved here for this to make a significant impact right away. All the autonomous development has been completely limited to ideal conditions. Once beyond that the vehicle has no choice but to let the driver take over.

Cars aren't even close to being able to handle a suddenly, completely snow covered road, and thus you can't sell something 100% autonomous in 75%+ of the country.

 

Let's pray then that this sad day when cars will be sold without steering wheel never comes! This will be a drama if it happens. Especially if it happens unilaterally, globally oriented by the insurance trusts or industries. We will have to organize a revolution then! Hell yeah!

 

All weather testing has been going on for a while. But if you look at the stats in term of car death with the expectation of a 90%+ reduction (these things should approach elevators). Right now in the US we are running around 35K car deaths a year. That's a lot of mothers, fathers, and kids that we lost for nothing. And, often, it isn't the person making the mistake who dies. Don't get me wrong, I'll miss driving, but eliminating the chance someone texting and driving kills a kid, or someone's parent, its hard for me to want to stop something that'll fix that. Good news though, VR racing will be a ton better by then, and drone racing already is a ton of fun (and far cheaper if you crash than in one of our cars). Granted you need wicked twitch skills, which I no longer have.

Oh, and you folks do realize the human carrying drone trials in Nevada started last year and Dubai is expected to have the related service in production next year. Now I could really get behind flying to work though, all of those heavy drones flying over my head does give me pause. But, if it isn't people, it sure will be Amazon packages shortly.

 

Thank heaven for third world countries where they will be driving old beaters for the next century or two. Costa Rica is quite nice, and unlikely to be going ape$hit with driverless technology. You think computer hacking is a problem now?...wait 'til the ne'erdowells start mucking with car to car communications.

 

I respect your research and knowledge as an analyst; but I don't think you understand how little cold weather testing has been done in the industry. And the cold weather testing that has been done has been mostly a failure. There's a reason every "rolled out" autonomous system is in a place where it never snows.... Nevada, Dubai, etc.


another thing to look into, which I imagine you have... is the studies that have been done about our desire for fully autonomous cars. For every person who would buy one, there's someone who absolutely wouldn't. There are millions out there who would rather be in full control than die because a machine made a mistake 0.0001% of the time, even if the chances are higher. This will only get worse if cars are given the ability to kill the driver if it spares pedestrians or alike.

 

I think the debate ends after Stohlens last intervention! Great finale man!

 

Google’s self-driving car has indeed had issues the snow. Go figure a company that largely did its initial testing in draught ridden California would have issues with weather? But their effort is largely for in city and really is much like an elevator with wheels. Some of their initial designs looked more like rolling living rooms than vehicles. Audi and Ford have been testing in weather for some time and the tech industry has had a number of solutions, like headlights that can almost magically pierce snow and rain, for some time (which Jaguar is still expected to deploy first by the way).

It is kind of amazing how coverage of self-driving cars tends to bifurcate. That is one of the reasons one of the models being used is horses vs. cars. If you think about it, when we had horses we kind of had self-driving. You could sleep o one and still largely be safe. Back when cars came to market folks argued that they were unsafe (they were), unreliable (they were), expensive (they were), and anything but elegant. Yet the outcome was nearly absolute. Back then news and advancement were relatively slow, initially cars were built one by one, not off assembly lines, and collaboration between competing firms non-existent. Inou about 10 years the computer to control a self- driving car has gone form one off monsters that barely leave room for one passenger/driver in a 4 seat car to a small computer module that is smaller than many desktop computers and prototypes that now approaching laptop size.

Now 35K annual deaths means around 100 people die a day in car accidents, once we get to critical mass this number will drop a lot but that will raise the focus on each individual death far more and one human driver taking out a bus, like just happened near me, should raise to international news levels sometime next decade. That’s the kind of event we are anticipating to force human drivers off the road.

You are right, a lot of folks will fight this, but we are currently seeing an increase in accidents by distracted drivers and a new variety of drugs. We have testing that shows that when placed with autonomous cars a significant number of car drivers get more aggressive and given these cars are only truly vulnerable to humans how misact it isn’t much of a jump to see that, unless that changes, the problem that governments will fix will be human drivers.

 

I believe the future enderle is discussing will come. I think 2050 might be realistic. In 2030 I expect we'll still see many human driven cars still being sold and a surprising amount of internal combustion engines.

 

I'm not arguing that autonomous vehicles are safer, as you keep pointing out, that's obviously true. I'm arguing your timeline.

No one has an answer for bad weather... no one. The car is simply not smart enough yet to handle a snow covered road. Humans will need to be paying attention for many years still. We're developing 2020 models now, so you can't convince me that by 2030 a majority of the NAFTA market will be 100% autonomous. It's not gonna happen, we don't move that fast. Not to mention it takes 20 years for the current fleet of vehicles to cycle out anyways. 13 years from now there will be plenty of current vehicles on the road and those people won't be able to afford a new technology car.

 

It is a bit confounding that society is willing to spend billions on the development of driverless cars to save a mere 35k lives per year when there are so many diseases, each killing far more people annually, that are going without research funding. I would gladly give up the airbags in my car to find a resolution for Alzheimers or Parkinson's.

 

One doesn't preclude the other.

 

Agreed. ......I'm old enough to remember that in the '60's and '70's everyone thought that in the year 2000, we'd all be driving hover cars and living in bubbles......instead we live in those same houses and have new cars that look like 1969 Camaros, 1970 Challengers and 1968 Mustangs! LOL


Dave