Some people suck at driving. Some are very good at it. Not everyone is in the lower 50%.
Self-driving cars will be the law when EVERY car is required to be. And that is because some human cannot be trusted with their own cars not to fuck with self-driving cars.
An alternative would be to have separate (but equal) streets for self-driving cars only. Like, a floor above ground level. Won’t need every street with its own second level.
Heck, I’d go one step further (into the future) and have a system of self-driving cars that pick up people at stations, like one-passenger/family buses. You pay, you go, you get out, you forget about that car, you get another one when you need one. who owns them? Some corporation. Maybe you own stock in the companies.
I learned to parallel park using my mom’s suburban and then took the test in my grandma’s sedan. That was the easiest part of the test because the space allowed would have fit the suburban. The hardest part of the test was not california rolling through stop signs my dad kept drilling it into my head that the car had to rock back for it to be considered a stop… As in do what I say, not as I do.
Back when I got my license there was a written test to get your learners and then a driving test for the full. Because I was rural, everyone had plenty of experience driving before they hit 16. Tractors, pickups, and some back road no license driving. It was a right of passage, on your 16th birthday you wrote your learners, then immediately booked your road test, and went home fully licensed.
Nowadays you have to have your learners for a year or so.
I used a stick, too. I knew my dad was shopping for a new car. I hoped he would switch to an automatic. He came home, I asked, he said he got another stick, then he said “let’s go take your driving test today”. I passed.
All of this, except we weren’t rural enough for tractors, but my grandpa used to pay me $20 to mow their 5 acres on the riding mower. Was a good way to get a tan and practice turning where I couldn’t hurt much.
Also, my dad encouraged poor decision-making and had me driving to the grocery store by myself for every little thing by the time I was 13/14. I was happy as a clam getting to do that but man could it have screwed some crap up if I’d gotten pulled over. Pretty sure I only learned how to drive stick because he wanted to send me to the store rather than having to go himself. My mom’s vehicles were all automatic transmissions by then.
In IL, min of 9 months . . . with zero citations on your record during that time.
So, if you’ve had the permit for 8 months, get a speeding ticket that you can’t negotiate down to a non-moving violation; you’ve got another 9 months of being on the permit (until age 18) before you can try to get your license.
Another current example of a Tesla failing to handle downtown Oakland.
At various points it: slows to a crawl, stops in the middle of intersections, drives on the double yellow, misses a “no turn on red” sign, ends up on the left side of the road at one point… altogether about as bad as my own worst attempts at city driving.
I know Tesla isn’t the only one developing autonomous vehicles right now, but they are (afaik) the only ones willing to let you see how bad they are.
I’m just thinking here. These are my own thoughts and predictions. It’s not a frequency issue. It’s a severity issue.
Let’s say that it is possible that self driving cars can be made so that traffic fatalities fall below today’s levels, as measured by annual US traffic fatalities. According to wikipedia, that is a number like 36,000 life per year (pre-covid).
Now let’s say that on average currently, this creates a liability of $500,000 per death that is borne by the driving population, which is effectively financed by the insurance industry in return for premium. That is 18B in compensation to accident victim fatalities on behalf of the at-fault drivers.
Now, you imply that roads would be safer with super-computer controlled cars.
That remains to be proven, but let’s just assume for a minute that it is true. Let’s say that we lower that to 25% of current fatality rates, or 9,000 per year. You argue that this is a more optimal situation so that it must come to fruition.
I think that the cost of a fatality will rise from 500k to 10M per death. It’s not John Q Public as the driver at fault anymore. It’s deep pocketed Tesla or GM or Google who is “at fault” and I use the quotation marks intentionally to highlight the concept that plaintiff attorneys are known to convince juries of fault where very little actual responsibility exists. In P&C insurance, we call this social inflation. This could be the biggest trigger of social inflation that our tort system will ever see.
Now the system has to finance 90B for fatalities, and Tesla and Google lose interest in the technology. Sure, the cost of personal auto insurance plummets or even goes away. But the price of the product liability will skyrocket, and must be borne by consumers.
It’s not the lack of technology that prevents the adoption of driverless cars, it’s the plaintiffs’ bar.
Game theory teaches us that a system can and will settle into a non-optimal state.
I disagree with your prophecy. I think perhaps you’re right that self-driving cars will drive up the cost of fatalities, but in the long run that cost will be applied to both AVs and non-AVs. This will make AV tech more valuable, not less.
OTOH, I do think something like “social inflation” is putting AV development in a “non-optimal” state right now.
Like, looking at my video above, it really doesn’t make much sense to have non-professional teenagers stress testing (bad) self-driving in downtown Oakland. It would make much more practical sense to perfect pizza delivery in the suburbs, or trucking on the highway. But ironically, letting dangerous customers use their dangerous technology in dangerous places, is how Tesla minimizes liability.
Also, just last week, Waymo released a paper about accidents that they supposedly would have avoided. Which is pretty silly when you think about it. They are obviously desperate to equate lives-saved to lives-lost.
In the long run though, I think we will equate lives-saved to lives-lost. In part because there’s no good way to distinguish between AV, non-AV, and semi-AV.
But yeah, I think if self-driving vehicles are only 4x better than human drivers that won’t be sufficient to motivate people to switch. And the liability to the manufacturers will be unacceptably high.
I think they’re going to have to be more like 50-100x better than human drivers before they really take off. And considering that they’re currently worse, that’s a long ways into the future.
I certainly don’t think I will ever own a Level 5 autonomous vehicle, and I wonder if my 5yo niece ever will.
I think you’re underestimating the ability for technology to get underneath your fingernails without anyone noticing it.
I can’t say if the tech will ever truly mature, but I bet good odds that you will…
order products that are partially delivered by (level-3 / level-4) AVs.
buy a car that can drive you from point A to point B, even if you twig decide to drive to point B yourself.
IMO, the main thing that drives adoption of new technology is not that anyone actually wants it, it’s that it’s that the marginal cost of production is small.
Lots of opportunity. Eat breakfast on the way to work. Start a long drive late at night and sleep most of the way. Arrive refreshed. Lots of people are nervous drivers on highways and in cities, this eliminates that possibly.
And maybe some peer pressure once it becomes common knowledge that people driving cause bumper to bumper traffic and slowdowns where self-driving cars will limit or eliminate that. Nobody enjoys the morning traffic, if the slowdowns are perceived to be that one guy who’s hands-on driving, maybe that pushes people.
This is so overlooked. The amount of traffic eliminated by an AV only mandate will be significant, if not completely. Once people realize there will no longer be rush hours, metropolises will convert in unison.
I dunno, Americans aren’t exactly known for doing their individual part in order to enjoy a better communal situation. You’ll have half a country worth of free riders.
This doesn’t make sense to me. What’s the logic? How will the means of conveyance affect the number of [strike]drivers[/strike] passengers that want to get someplace?
People suck at driving (can’t stress this enough). Experiments have been done where even without any reason for traffic (people driving in circles), traffic will happen because of human drivers. One car slowing down would cause a chain reaction, eventually leading to a car coming to a complete stop.
If all cars on the road are AV, there will be no need for traffic lights, as every car will know every other car’s velocity and acceleration, driving with precision without ever having to come to a complete stop, eliminating nonsensical traffic.
my dad kept drilling it into my head that the car had to rock back for it to be considered a stop… As in do what I say, not as I do.
good luck with that. The beatings will continue until morale improves 