Tesla announced last night that all of its cars now being produced are fully equipped with all the hardware (sensors, cameras, and a 40x more powerful processor) to allow them to operate fully autonomous (L5) once government approval has been secured, via a downloaded software upgrade. Tesla says the software to implement “Enhanced Autopilot” ($5,000 upgrade), which offers a number of enhanced features,will be ready to download by December 2016, subject to regulatory approval. When the “Full Self-Driving” ($8,000) will be validated and approved is as yet unknown, but 2018 is being suggested.
In a PR exercise, Tesla will be sending one of its cars drive fully autonomously coast to coast shortly. This video of a Tesla operating fully autonomously (in a rather benign environment) gives a good idea of what that experience will be like; obviously the “driver” in the video still needs to be alert and have his hands at the ready, due to current regulations.
My son Edward (Tweetermeyer, on Twitter, and other outlets) has been following both Tesla and AV (Autonomous Vehicles) very closely for some time, and has teamed up with Alex Roy and Damon Lavrinc on a weekly podcast (“Autonocast”) to discuss the rapid changes in the autonomous field. I listened to the latest one, #3, during my lunch yesterday; it’s engaging and a great way to keep up with the rapid changes. It’s embedded below the jump. It was recorded two days before this latest announcement by Tesla, but there’s still lots of good stuff there, and gives an excellent overview of the rapid changing landscape of AV. Update: there will be a fresh one tomorrow (Friday) in response to Tesla’s latest move.
A song from 1966 to accompany a video of a car for 2020. Gotta love the Rolling Stones. Paint it Black was a good choice but Black Limousine would have fit the subject matter even better.
The idea of a software update to fully automate a car is frightening. My computer “crashes” once in a while after updates. Not fun, but not deadly ?
Manufacturers have been updating the software computers on their cars for years now. And Tesla has been making a number of updates for some time. This is how it is now. There are safeguards, and self-diagnostics. It’s not quite the same as updating a Windows PC. 🙂
The obligatory “Skynet” reference is of course not necessary.
Or this obscure reference…
Dave Bowman: Hello, HAL. Do you read me, HAL?
HAL: Affirmative, Dave. I read you.
Dave Bowman: Open the pod bay doors, HAL.
HAL: I’m sorry, Dave. I’m afraid I can’t do that.
Dave Bowman: What’s the problem?
HAL: I think you know what the problem is just as well as I do.
Dave Bowman: What are you talking about, HAL?
HAL: This mission is too important for me to allow you to jeopardize it.
Dave Bowman: I don’t know what you’re talking about, HAL.
HAL: I know that you and Frank were planning to disconnect me, and I’m afraid that’s something I cannot allow to happen.
+1
+2
+2001 ?
“Our compliments to the M-5 unit. And regards to Captain Dunsel. Wesley out.”
CC Effect… sorta. Spotted a BMW M5 sporting another TV reference yesterday, his vanity plates said “MACH 5” (Speed Racer)….
…but what went thru my head was Dr. Richard Daystrom saying, “M-5 tie in.”
Herbert… Herbert…. Herbert….
One of the most famous sci-fi films of all time (perhaps second only to Star Wars) is now “obscure”?
There is a huge amount of money being invested in self-driving cars. I am really not sure why the big push, for something that I am not sure that people even want very much.
Also, It can only increase the cost of cars, which is too high as it is, in my opinion.
I would be interested in hearing what others think of all this.
Not to mention the mass unemployment since fleet operators’ll go autonomous long before personal-car buyers,
Actually, cars today are about 25% cheaper to own and operate than in 1968: https://www.curbsideclassic.com/blog/operating-costs/the-cost-to-own-and-operate-a-car-has-fallen-25-since-1968/
Maybe you won’t want it, but many will. Commuters will be happy not to have to endure the daily grind. And autonomous taxis will be a great business opportunity, and will lower their cost. And the cost of this technology will tumble with time, like all new technology. ABS brakes, air bags, stability control, etc. were all quite expensive at first. Now they’re standard on the cheapest cars.
The upside will be transportation sitting in the driveway when one is no longer to safely drive due to age or impairment, assuming the driver as a fail safe/backup requirement eventually goes away.
I would imagine eventually insurance rates will go sky high for those who choose to drive themselves.
The cars should be designed to not operate if any fault is detected in it’s operating system and components. In a perfect would all collisions would be no fault/no liability to the occupants.
We’re dealing with my Dad’s advancing dementia, and after an episode where he never showed up for lunch at his caregiver’s home (two blocks from his house), I used his “help button” tracker and located him 40 miles from home. Thankfully he must have asked for help, as I watched his location reverse and he got home safely, upon which he lost his car keys for good.
While I personally have little interest in autonomous cars, I could see where this would be a good solution for dementia patients or others who can no longer drive – the vehicle could be programmed to take them to certain approved locations, for example. That would at least keep some level of freedom of mobility, for a time, any way.
I remember getting a call from Dad (he had Parkinson’s) at work. He had fallen and was bleeding, Mom was at the store, he was trying to call her but got no answer. As I pulled into the neighborhood I saw their ’93 LeSabre a few blocks from the house, wandering around aimlessly and driving by the house a couple of times.So I caught up to see if they were both in the car.
Blowing the horn and pulling up next to the car failed to get her attention. I finally passed her and as I pulled up to a stop sign, got out and found Mom alone, with her cell phone ringing away as Dad tried to call her. It took her a moment to realize who I was, she told me she was going home. She had been driving around the neighborhood for about an hour. Alzheimer’s is a very sad way to go.
After I got Dad to the hospital and patched up, I got them into assisted living. A few months later Dad had one last drive, he had hidden his still valid drivers license, told me he lost it and was mad I wouldn’t take him to be retested. I had taken the Buick away at the same time they entered assisted living. The poor Buick had a new dents in it everywhere by this time.
He took a cab to a car rental place and managed to get a rental, he could hardly walk or talk so I don’t know why the stupid person at the counter thought renting him a car was a good idea. Only found out he left the assisted living home later when they finally called saying he was missing, and my Mom uttered a few words about a car rental. She had gone with him but made him bring her back to the home, which he did and then drove off.
That night around midnight the home called, he was back with 3 different cars following him, all with stories of how they were almost hit or run off the road. The Chevy HHR rental was almost out of gas, filthy and bug splattered, and had over 400 miles put on it. The seat was urine soaked. There wasn’t a scratch or dent on it.
This time I found the license and took it away.
Autonomous car would have been just what was needed for them during their final years.
Yes that sounds good on paper. But what happens when they exit the vehicles at that location? Are they going to know where they are or why they went there in the first place? There would almost need to be a tracker on that person tied to the car in question. I wouldn’t have wanted my dad to go someplace in one of these vehicles only to find him 5 blocks away from the vehicle and hour later.
I completely agree on both points. As far as the self driving technology, maybe I’m as outdated as the DIY cars we currently use.
As far as the cost; ok a Tesla is going to set you back a fair amount of change regardless but in general it does seem that cars (new) are going to become far to expensive for most people to afford very soon.
Considering the sociopolitical ramifications of this I’m not sure I feel discussing this here will go over well, but when business opportunity(for business owners) rears it’s head, I feel that’s your answer as to why this is being pursued. Meanwhile us drones are being sold like it’s a feature akin to cupholders “oh how convenient!”.
As a long distance commuter in the SF Bay Area, the self-driving car cannot come soon enough. My vision is a Sprinter luxury passenger van modified to have a sleeping area in the back, small eating area and living area in the front. I might go to bed in it and just let it take me to work at a designated time. Then it wont really matter how long my commute is. It can go Uber people form the airport while I’m at work. I may not even need a house.
I’m only half-joking
added self-driving bonus: vehicles will never stop to rubberneck at accidents on the hihgway
Great….
It is, actually. I can’t wait to have 80-90% of the drivers out there, who suck at that activity, be riding in AVs, where they belong. I’ll be a lot less worried about some idiot crossing the center line while texting, eating or jacking off, or just falling asleep behind the wheel.
And it’ll be fun to mess with them, with certain maneuvers. 🙂
Well, when you put it that way, I’m sold lol
+1. A few years ago I would have been harumph-ing over this “nonsense”, but living in NYC for 10 years had a dramatic influence on my views about driving in the 21st century. My main exposure to contemporary traffic and driving habits since 2007 has been from the back seat of livery cars in Manhattan and Brooklyn, and from that perspective in an environment where drivers MUST be on their toes and 120% attentive (and believe it or not, they usually are, for the most part) things didn’t seem like they’d changed so much.
Cut to late 2015 when I moved to a fairly suburban area of Florida with average traffic patterns, a skew of average drivers on any given day (varied ages, experience levels, degrees of training, etc) and a fairly benign driving environment with good roads, minimal foul weather, good signage and nary a hill nor blind curve. My reintroduction to daily driving was one hell of a smack to the gob, for sure. Yes, my time off the road probably increased my wariness once I dove back in, but it’s clearly not my imagination telling me that folks are really not giving serious consideration to the concept of focusing on the task at hand.
Personally, I balk at driving a vehicle with an automatic transmission, let alone one that takes over any other task that I feel should be under my control, but I’m impressed as hell with this technology. I’m watching and waiting anxiously!
The “accident” that wrote off my Outback would almost certainly not have occurred if the F150 that hit us was equipped with this technology. The technology spooks me a bit but Paul is basically correct about most likely less things to worry about.
https://www.curbsideclassic.com/uncategorized/coal-2014-subaru-outback-3-6r-limited-oh-we-hardly-knew-ye/
I so agree with Paul. I saw a bimbo, on I-5 North near The Curves, steering with her knees, eating something that looked like a burrito, while putting on make up, probably because she was already late for work after spending the night at her ex-boyfriends house trying make up sex. They ARE out there. And I do not want my 33 year old Ranger taken out by a nitwit like that.
Paul,
I think you’re missing the elephant in the room – you’re assuming that the 80-90% of drivers who suck at driving will be in AV’s but that you’ll still be tooling along in a car you enjoy driving.
Ain’t gonna happen.
When 80-90% of people are in AV’s – or for that matter, when 30-40% of people are in AV’s – then no matter how good of a driver you are, you will be the hazard because a) you cannot drive as well as a computer, and b) even if you can, when accidents happen, the driven car will always be viewed as at fault since the AV couldn’t possibly have been in the wrong – there’s too much money at stake to let that happen.
The only way for AV’s to reach their potential – market or otherwise – is to completely eliminate the human element – i.e. – get human driven cars off the road entirely. So whether you are a skillful driver or not will simply not matter. You are human, and thus cannot be trusted with a two-ton human operated projectile. You choice to drive yourself would pose too great of a risk to the substantial numbers of other people who are wealthy enough to have purchased a new AV so they don’t have to drive, and who will be on the road reading their iPads while being shuttled to wherever they’re going.
So there will not be 80-90% of people on the road in AV’s. There will be, by law, 100% of the people on the road in AV’s. And you will be one of them, or you will not be on the road at all.
OK; that was just me projecting my own wishes, but it’s not really how it will likely unfold on the shorter-medium term.
Yes, in order for AVs to reach their full potential, one would want to get rid of human drivers. But again, I don’t see yet how and why that would happen. There are way to many places/situations where that’s not going to happen, mandating everyone to be in AVs. What about the folks that live in out of the way boonies, and on ranches, and other obscure locales. I suppose we could say it’s theoretically possible, but not realistic.
And are folks who have classic/older cars going to be denied driving them? Their political clout is considerable; there have repeatedly been efforts to get them off the roads because of their emissions, but it’s never happened.
Realistically, AVs will start in big metro areas, for those that want them and for businesses that can take advantage of them. I really don’t want to speculate how it’s going to unfold in the long term.
Some folks still have land line phones. Are they going to be forced into cell/internet phones? I’m not sure it’s comparable.
Are we going to ban cash someday? Require folks to have some kind of personal digital device? I don’t know, but I rather doubt it.
In many places of the country, especially smaller towns, rural areas, and in the mountainous and snowy areas, AVs may not be even be very feasible, let alone desirable.
Some folks kept their horse and buggies for quite a while. Some farmers kept their mules for quite a while. Some still use them today. But will folks eventually find AVs attractive and advantageous? And will society create incentives for their wider adoption? Certainly possible.
Thanks, Paul, for being a voice of reason.
However, as promising as I think AVs may be for long trips and certain commutes, I’m assuming they’ll never be foolproof nor accident-proof, as I’m afraid some expect. I can’t imagine that any device could possibly guarantee absolute flawless driving in, say, severe winter weather. And of course, when things such as deer dart in front of a vehicle, all bets are off. Accidents will never completely go away.
That being said, I’d wager that a properly sorted out computer algorithm can handily outperform most drivers in emergency situations. As for those who argue that technology fails, well, so do human drivers.
Paul,
Last week (I think it was last week) the NHTSA released some long range goals. Their stated objective is to have 0 highway traffic deaths within 30 years.
There’s your timeline. It’s a goal I think cannot be achieved, but it certainly cannot be achieved if humans are allowed to continue to drive their own cars.
Now, obviously, there would have to be some societal changes for people to just give up their cars, but look at the societal changes in the last 30 years. Had you been told in 1986 that the US Govt. would be listening in on its own citizens or reading their mail (e-mail) or making them go through intrusive full body searches to board a plane as is done today, and that while some may complain, most are grudgingly accepting of the intrusion, would you have believed that?
Paul,
In my opinion, you’re both right and wrong.
As you say, the technology might not be able to work in some remote places or because of climatic conditions.
But it will work in big cities or main roads and highways. There, the driven cars might be considered to be a liability to AV passengers.
So it’s highly probable that driven cars might be forbidden there. In other words, driven cars might not have an access to cities, highways and main roads.
I can’t see why it shouldn’t happen. In Europe, we’ve got low emission zones to fight against air pollution. One was created in Paris the 1st of july 2016. Since that day, every car prior to january 1997 and motorbikes prior to june 1999 is forbidden in Paris during weekdays, between 8 AM and 8 PM.
The same legal system could easily apply to driven cars to regulate their access in areas where only AVs are allowed.
Moreover, in Paris, there is an exemption for any 30 yo + car with collector plates on the belief that they do not run very often and does not contribute much to air pollution.
You could have the same kind of exemption in areas forbidden to driven cars on the behalf that classic cars aren’t used very often and, as a consequence, aren’t such a liability.
I’m not saying it will happen, but I strongly think it will someday. At least in Europe since the idea of regulating the access of certain cars on the basis of their age and/or technology already exists over here.
Moreover, if the regulator doesn’t forbid driven cars, insurance companies will if they consider driven cars are a liability. Insurance premiums might skyrocket in a way most people won’t be able to afford a driven car anymore.
In France, I heard that insurance companies are considering offering lower premiums to drivers who accept the equivalent of a flight recorder on their cars.
The technology is almost here and the legal framework is already here so it might be a question of time before driven cars are forbidden in some areas.
“In France, I heard that insurance companies are considering offering lower premiums to drivers who accept the equivalent of a flight recorder on their cars.”
To some degree, this is already done by a few insurance companies in the US, through a device plugged into the vehicle’s OBDII port. Progressive Insurance’s Snapshot was one of the first, and Allstate offers DriveWise. There are others (perhaps State Farm and Liberty Mutual, as well).
Boeing’s have been capable of autonomous flight, gate to gate for quite a while now. None of that capability is used, except in rare emergencies, in the Untied States. Auto landing in zero visibility has been used in Europe for decades. US regulatory, public attitudes and tort legal system all stand in the way. I see similar hurdles for large commercial trucks. Personal cars may come first.
I think there are still some thorny issues around liability. Who pays if my autonomous car crashes? The manufacturers will fight tooth and nail to avoid that liability.
I think that, instead, the autonomous cars will be owned by companies like Uber, who can go toe-to-toe with the manufacturers. Furthermore, the liability of any one snafu will be spread across a whole fleet, which will be, overall, safer than human drivers.
I know that, personally, the benefits to not driving are not having to worry about parking and not having to worry that splitting a bottle of wine over dinner will make me a felon. Those benefits accrue only if I don’t actually own the car. 🙂
Who pays if my autonomous car crashes?
Your insurance, unless the AV software/system can be found to have been negligent in its design. As Musk puts it: “who is at fault when an elevator fails?” Are you going to sue Otis? Well, if the design is truly faulty, yes. Otherwise, no. Shit happens. That’s what insurance is for.
I think that, instead, the autonomous cars will be owned by companies like Uber,
Well, with this announcement, that’s clearly not so. Every new Tesla will be fully AV capable. And Tesla will have its own ride-sharing network. BTW, Tesla’s AV system will not be usable for use as an Uber car, to protect its own competing service. So Tesla is already years ahead of Uber.
In most states, liability insurance only pays when the insured driver is at fault. Run a red light because you were careless? Insurance pays. Run a red light because the brakes failed with no warning on a well-maintained vehicle? You are not negligent and your insurance will defend you in court.
A driverless car runs a red light. Is the guy behind the wheel possibly negligent? Maybe in some cases, but maybe not in most. So is it the responsibility of the manufacturer (products liability case)? You can bet that they are going to fight that. Was it the fault of poor maintenance or the owner not getting software updates done? I see a minefield here. Or do you just lease or rent the car and the owner maintains products liability coverage on it. I would see no reason for a driver to maintain traditional liability insurance for negligence when he isn’t driving.
No-Fault insurance was a fad of the early 70s that has stuck around in some places, but perhaps this would be where it might actually make sense. This is going to require some massive re-thinking of auto insurance.
Good points. Undoubtedly, insurance will have to change. But it’s pretty safe to project that it will be lower for whoever does pay for AVs. There’s pretty strong consensus on that issue. But it may be a challenging transition to work out the wrinkles.
“Who is at fault when the elevator fails?”
Maybe the company who made it. Maybe the company that maintains it. But certainly NOT the person who presses the button for the floor they want and is at the mercy of the machinery.
But certainly NOT the person who presses the button for the floor they want and is at the mercy of the machinery.
No one suggested that, right? I’m not sure what your point is.
As I said, the general consensus is that insurance will go down for AVs, perhaps quite dramatically once they establish a track record. But some kind of insurance scheme will still be needed. I don’t see it as a major issue though.
From the perspective of a 25-year veteran of the personal insurance industry: You’ll still be liable, your insurance will respond just as it does now, and as a bonus you’ll probably pay a whole lot less for that coverage than you are today. Airbags, anti-lock brakes, stability control and advanced anti-theft systems have lowered insurance rates over the decades in small ways that are hardly noticeable. Now consider a safety system that actually minimizes the influence of the one system in your automobile whose failure is responsible for 99% of all accidents: The Driver.
Just as you’re responsible when your dog bites someone, you’ll be responsible if your car hits someone. Even today if your handbrake fails and your car rolls into traffic causing injury or property damage you are responsible and your insurance will respond first. If that failure can be attributed to a design flaw or technical failure by the manufacturer the onus is on your insurance company to subrogate against that entity to get its money back.
From a personal liability (and sociological) perspective AVs are just what the average American has been dreaming of for years: Another way to shirk any personal responsibility.
“Even today if your handbrake fails and your car rolls into traffic causing injury or property damage you are responsible and your insurance will respond first.” – this is true *if* you are somehow negligent in the situation that you cite. Insurance is only there to pay for the stuff that you would otherwise have to pay for on your own, and you would only be responsible if you are negligent, whether at common law or by violation of some traffic statute.
I am not sure what state you are in, but where I am, I would defend that handbrake case all day long. If you have a modern, well maintained car, handbrakes don’t normally fail. Therefore, it is not reasonable to expect that it would or that you should not rely on it. There is no liability on the owner here, so the injured person is not compensated. If you have an old car in a rust state and have had prior trouble with the parking brake, they yes you are at fault. But where there is no liability, insurance does not pay. What liability should a passenger in a driverless car have? The law only requires reasonable care, so when I am doing what the owners manual tells me to do (which is to not touch anything and let the car drive itself) and all appears normal, how am I responsible when the car (without warning) swerves into oncoming traffic or onto a sidewalk full of kids? If I am not negligent, there is no recovery by the injured person.
In most states, its the car and its owner liable and insured, not the driver, per se. If something malfunctions, driver or the car, the owner is assumed liable and insurance pays. I see a potential difference here. In airplane crashes, everyone gets sued, owner-operator (airline), manufacturer, parts makers, repair stations, etc.
^^ Bingo. By virtue of the fact that you’ve opted to own a vehicle, you’re automatically assuming a degree of liability. Motor vehicle insurance law essentially states that once you take ownership of said motor vehicle you are responsible for it, so whether the owner is behind the wheel or not he will be the first target of liability, then once it’s proven that he couldn’t have prevented the problem that liability is shifted on to the next most viable party/entity.
I would say were going to see the model outside of the auto one where everyone gets sued at least for a while. Since most crashes are related to people not mechanical malfunction (when I did claims for our unit that was well under 1% of claims) I imagine there will be a big push back onto the industry for awhile.
There have been many discussions about the insurance. Volvo has claimed they will cover the costs if one of their cars crashes in auto mode. Right now they will be treated as safety systems, in the future it’s been discussed that liability will be shipped to the vehicle builder unless maintenance issues are expected. In the future I imagine if you have a full autonomous car with no manual mode you will have insurance that covers theft and property damage but liability will be an add on.
Great podcast, Paul!
It should be interesting to see how this affects very old drivers. I hope this means I can “drive” well into my second century if I can still afford to own a car. I’m probably too old for that kind of timing to work out, but many of our members might still have a self-driving 30 year old beater in their parking spot that can get them to town when they turn 100 years old.. So much for the kids taking their keys away…
Also, the idea of running over people wearing “Make America Great Hats” would be quite a plum for hackers both inside and outside the country. While the geo-system is learning about local conditions, it could be told the wrong info on purpose by people who want to cause “a disASter”.
You could also see the car getting smart enough that while it’s protecting itself and its occupants, it might choose to plow into a crowd to do so. You’ll want to be on the inside. New laws needed here.
Great podcast, Edward! actually. 🙂
And yes, this is a huge potential boon for older drivers.
And you could send your kids to school or lessons in a driverless AV. And you and your spouse could share one car; It takes you to work, then her, and then picks you both up ,jointly or separately. Takes you to dinner; no worry about how many drinks.
Yes, Edward is certainly well versed. He and his podmates make for compelling listening. I listened to both ‘casts, and their enthusiasm for the subject is clearly evident. It’s interesting to note that they say people actually on the ground doing development on this technology are often those who can’t or won’t drive. And, like most people who really like doing it, my knee jerk reaction is “NOOOO!”, but the longer I’m on the road, the more I wish there was a wall between me and the guy coming the other way. The only kind that can work is the one that the vehicles create. And here’s a vote from the side of me that likes to ride my motorcycle. No more accidents caused by people who “just didn’t see the guy on the bike”.
The more I ponder, the better it sounds. We have had several horrific local accidents in the last 5 years involving drunk drivers who entered the wron side of a divided highway and pressed pedal to the metal until they hit someone head on. The time for that possibility to end can’t come too soon.
I must also raise my hand as someone who has occasionally texted while driving. You would think that in my 6th decade I would be smarter than that. It’s unlawful and unsafe, and still, I look around and say, “Nobody around, I’ll just text one sentence.” It wouldn’t be dangerous for the other guy if I can’t hit him. Even using Siri, your eyes can be off the road.
Wonder how long it takes before self driving vehicles will be mandated, though. Will there be long stretches in the country that are marked for “personal driving” where you simply sign off, go through an electronic gate and take over the wheel?
So long as there is a provision that the car might tell you it is faced with conditions beyond its capabilities, and require the driver to take over, it would seem the car will ALWAYS require a competent and sober driver.
Will autonomous cars become so perfect that they never need human help? Maybe, just as soon as it can remember there is a mammoth chuckhole under the fresh snow on the corner of 90th and Fort. Or knows that the city just fixed it before the most recent snow, and does not hold up traffic waiting for a chance to swerve around it.
Personally, autonomy seems so complex and there are so many surrounding issues regarding its responsible use, I think there may be a movement back to the idea that certain roads may have embedded systems that efficiently create trains of unconnected cars that move in an efficiently spaced manner on major highways and thoroughfares. This would be an ideal solution for Southern California commuters.
But, when I get off the highway and travel sketchy gravel roads to my mother’s acreage and drive across the grounds to use my truck as a ladder in the orchard, I don’t see how a 100% autonomous car will ever figure out that one.
There will never ever be a 100% perfect autonomous system until non humans design it. We are flawed, the technology we make is flawed, it’s programming is flawed and there are so many variables such as how all this stuff ages, cost cutting certain components, road and weather conditions etc so that is just a reality.
Reminds me of a trip that me and friend took last Winter in his 2016 Accord V6 Touring sedan. The snow turned to ice and the front of the car was plastered with it rendering the adaptive cruise control completely useless. We call this feature a Summer Spring convenience.
It should be interesting to see how autonomous driving percolates through the automotive landscape in the coming years. This is going to be a game changer for the infirm/disabled, who will have much more options available to them in terms of mobility, especially in the suburbs.
As for myself, the idea that I could drive myself to the bar and have the car drive me back after a bunch of beers is extremely alluring.
The long-term mobility implications are fascinating – once fully-autonomous vehicles become the norm. It seems the impacts can go in just about any direction – on many important topics. Just a few that I’ve been thinking about:
Will AVs spell the end for public transportation? Local/state governments have proven themselves very poor at maintaining public transit infrastructure, and I can see them jumping at the chance to outsource the responsibility to private companies that operate AV fleets instead. I’m not saying that’s a great idea – just a possible outcome.
Will AVs improve congestion, or make it worse? The technology can enable high volumes of traffic to keep moving, but will that encourage more people to travel to the same place at the same times? Also, will AVs render the concept of central cities obsolete? Why bother with the high cost of real estate in a downtown if people can travel around quickly regardless of where a business is located?
Will AVs lead to the end of private car ownership… or will it make private ownership more attractive? It’s easy to see AV technology being used successfully by car-sharing companies to the point where many people won’t want to bother with owning a car any longer. On the other hand, the cost of a private AV might be very reasonable, plus they could go and park themselves – and no hassle of sharing.
And then there’s the role that government will play. The gov’t seems eager to assume a role in regulating AVs (which may be necessary to some extent), but that will give the gov’t the unfettered ability to track and control everyone’s movement at any time. Now, that’s some scary stuff.
Just a few random thoughts – and I’m sure there’s many others.
I can see autonomous vehicles actually INCREASING traffic.
People who don’t own a car for a variety of reasons – they can’t afford to do so, or are no longer healthy enough to drive – will now be able to call up an autonomous vehicle and use that mode of transportation. Instead of relying on mass transit, a taxi service, or even friends and family, they will call a service that provides an autonomous vehicle.
A few years ago, I recall reading that a car-sharing service was instituted in a college town. Contrary to original predictions, it led to an INCREASE in car use, because the people who used it were those who possessed a driver’s license, but couldn’t afford to keep a car on campus. The people who already owned cars, however, did not ditch them to use the car-sharing service.
As someone once said, all predictions are risky, especially those regarding the future.
I’m worried that without the ongoing labor cost of a driver, and with the medallion system already disrupted out of relevance if not existence, at some point the cost of putting as many autonomous cabs on the road as an operator can possibly afford will be less than the “cost” in lost market share of NOT doing so.
I think traffic related predictions in particular are risky. Has a promise of reduced congestion EVER actually delivered?
I never underestimate human nature to unravel the best laid plans: Give congested cities freeways, and we’ll move out to the burbs to max their capacity. Make cars fuel efficient for the long trips, and we’ll go out and buy the biggest tippiest ones available. Make them safe, we’ll go and use phones in them. Make phones hands free, and we’ll get smart phones that require touch and vision. Make the cars autonomous, and…..
Very well said on all points!
Yet fuel efficiency and safety overall have generally gone up. Yes, there are always unanticipated consequences or variables. But it’s the batting average that counts.
Has anyone ever promised reduced congestion? But ever-more sophisticated computerized traffic management systems are allowing freeways and roads to handle ever-higher vehicle loads. Congestion would be significantly worse without them. And I suspect that AV-specific lanes on the freeway, where they can close the gap between the cars and increase volume, will be a likely development.
AV specific lanes by widening the roads or AV specific lanes by cutting into the existing ones?
I’m not sure how well that’s going to work unless every autonomous car is identical. Bigger heavier ones are going to react to curves, road imperfections, wind gusts, ect. differently than smaller lighter ones, autonomous control can adjust to these variables faster than humans no doubt, but not enough I suspect to keep them mere inches apart without occasional contact. Not to mention the probability that AVs will ultimately infiltrate mass transit and at the very least trucking, and I just don’t see those lanes being brisk with them, unless they’re relegated into the luddite lane with me.
A very key part of AVs is their ability and need to communicate with other ones out there. That very much opens the possibility of them being “hooked together” in a virtual train. If anything happens to any one car, all of the cars in the trains respond instantly and simultaneously, as is most appropriate.
This is just one step above adaptive cruise control, and a key element to eliminate/reduce rear end collisions.
“The gov’t seems eager to assume a role in regulating AVs (which may be necessary to some extent), but that will give the gov’t the unfettered ability to track and control everyone’s movement at any time. Now, that’s some scary stuff.”
True Dat! The only (and it’s a big one) downside to this technology is the thought that we might soon live in a world where we’re no longer ABLE to control how, when, or where our machines work for us, whether legally or not. The Model T may have put America on the road, with all of the ills associated with that, but AVs can ultimately mean that our wanderlust could effectively be shut down by government or corporate edict, leaving us without even the option of being disobedient, as the machines we once lorded our will over will turn the tables on us.
No, the government is not interested in collecting the direct data from these cars, which would be mammothly expensive anyway. Each system/manufacturer will collect its own data to constantly improve their system. The government has already said what they will do: set standards. That’s totally different than monitor the actual cars and data. These are proprietary systems, and the govt. will not have access to them. They can’t access On Star or the equivalent numerous systems now, right?
The only possibility I can envision is some kind of intervention in an emergency situation, to stop a car being used in a crime or such. But that would still have to be fed through the system, not directly by them.
Realistically, no one would buy into a system that the government actually ran.
“No, the government is not interested in collecting the direct data from these cars, which would be mammothly expensive anyway.”
As someone who works in the collection and analysis of Big Data, I can tell you that modern computing power and analysis techniques make it not very expensive at all.
Industry is already very good – and getting better daily – at collecting data on your internet habits and suggesting ads based on those habits. The evidence is on this very site, for those not blocking ads.
What can be uncovered about a person via the data trails they generate these days is pretty scary, though I personally am much more concerned what credit rating agencies and employers are gathering than government.
Those who choose to publish racist or otherwise objectionable comments on social media, beware; employers are already paying companies to mine this data for prospective employees.
Joe, you need to read up on how these systems work. They’re constantly mapping their surroundings via multiple sensors/cameras, and sending all that back to the system central. Some very knowledgeable folks have estimated that it may well take a whole server for each tesla out there; and that total IT processing/bandwidth costs per year to tesla are going to very significant, like maybe several thousand per year. tesla hopes to recoup that by selling the higher level of software ($8k). It’s many orders of magnitude more data than what comes of cells or PCs. In fact, managing that data, which is vital to improving the system, will be a massive investment and a big drag on the finances of those companies that are going to do it (Google, Uber, tesla etc.)
You think that the govt. is going to spend that kind of money per car? Not. And why, exactly? Tesla is not going to just hand over info to the govt. unless it’s just certain metrics. See: Apple.
Anyway, I have little time or interest in tin foil hat debates. Sorry. Frankly, i don’t really carry who knows where I drive. If it’s that interesting to someone, help yourself.
Umm we are constantly tracked by our cell phones already. But really yes this is just more data that companies will have, not sure how that’s any better then the govt having it. Both scare me actually. A number of governments already collect huge amounts of data. Public security cameras police car mounted cameras, EZpass license readers at toll booths. Some towns have installed random license reading cameras around town (for road traffic gathering) but they track the plate numbers as well.
Unfortunately at this point everyone’s so used to it I don’t think there’s a way to roll it back.
I’ll report on my first ride on the autonomous Bustang when it happens. I believe the driver is currently the biggest continuing expense.
“And then there’s the role that government will play. The gov’t seems eager to assume a role in regulating AVs (which may be necessary to some extent), but that will give the gov’t the unfettered ability to track and control everyone’s movement at any time. Now, that’s some scary stuff.”
I’m pretty sure the government can already track and follow you if they want to badly enough. They can also affix a device to your current car to help them do so. They could also affix a “marker” and track you via satellite. Maybe “they” already do.
But if you’re really worried about cars I don’t see what’s stopping anyonefrom getting out of it and walking instead. Or riding a bike. Or strapping on some rollerskates. But don’t forget to wrap yourself in tinfoil from head to toe so you’ll be invisible to the overlords… 🙂
I have no idea if the government “wants” to regulate AV’s. I think they SHOULD so that there is at least a chance of some commonality of systems and the protocols used within.
Let’s just say that I think that the government’s role in an AV infrastructure will be far bigger and far more involved than you do. But I hope you’re right!
My fear is that the government will impose restrictions to “solve” some of the problems that AVs may generate (increased congestion, etc.) in the future. But let’s check back with each other in 25 years to if that’s true.
In the meantime, where’s that tinfoil hat of mine??
As cynical as I am about AVs, the whole big brother government tinfoil hat aspect really doesn’t worry me (not that it matters, as I’ll NEVER buy one! Ahah!). I think the role government will ultimately play is heavy handed taxation, presuming these cars will all be electric, and presuming speed traps and traffic enforcement as a whole will be a thing of the past, money has got to come from somewhere.
Fools.
That’s exactly what horse and buggy drivers said about the early automobiles. 🙂
Thoughtful discussion on the podcast about what should prove a very complex issue, both in engineering and legal and policy areas.
I wouldn’t be surprised if I could write a list of 100 scenarios or issues, several of which are opposed to each other in their nature – and I’m not an expert on any of this.
For the car to drive itself, it has to have a destination. If one simply wants to go for a random drive through some region for sight seeing, how does this work?
Well, you may have found chink in the system. 🙂
But it would be very easy to add a bit of software to accomplish just that. And quite likely, it will take you to/by businesses that are willing to pay for the pleasure of seeing your car come in/by.
You could drive the car yourself, or, when there is a fork in the road the car could ask which way.
Whether you are for or against this inevitable march toward machine control, there is one clear winner: lawyers. Patent lawyers and personal injury lawyers should be absolutely thrilled. Things will be messy for a while as everything gets sorted out.
To that end, one of the Tesla owners I know is a highly successful personal injury attorney (and a total prick). He purposely got the Model S so he can better “understand” what it does and how it might “impact” potential clients.
I like to drive. I think I am good at it. Unlike many today, I pay attention to the important task at hand behind the wheel… DRIVING!
While I understand they want to move forward with this amazing technology for safety reasons, because so many folks are not responsible behind the wheel, I for one do not want something I genuinely love to do being legislated out of existence.
I fear this is a slippery slope which we are about to go down.
I realize this is a bit of well written Curbside Fiction, but please click the link below and read this amazing story. I for one don’t want to end up like this guy….
https://www.curbsideclassic.com/curbside-fiction/curbside-fiction-the-future-is-now/
Thus causing a generation that can’t effectively read, write ,do arithmetic (or any mathematics), critically think, or drive without the use of some kind of machine.
I know it’s negative thinking but it follows current trends.
….sorry, just throwing it out there…
Well at least with those qualifications they’ll have lots of low skill jobs available.
Oh wait…
I have some reservations about this system, but then I am over 50. Is there anyone here who has not had an electronic device freeze or crash? The question is not whether the system will fail, but how often and in what ways. I don’t think it is reasonable to expect that none of these cars will suffer a strange electrical fault or mechanical failure – just think of every car you have ever had. Most don’t, but most isn’t all.
Human nature being what it is, when the system has worked reliably for the last 50 times I used it, I am *much* less likely to sit there with my hands at the ready, scanning every approaching hazard as if I were actually driving. No, I will be lulled into checking my email and writing a comment on CC. And when the car suddenly veers towards the oak tree or does not stop for the kid in the crosswalk, it will surprise the shit out of me, and I will probably not have time to realize what is happening so I can manually recover.
That said, it might work out that the system, avoids so many other human-error type of accidents that it would be a good tradeoff. Maybe I am foolish to think that I can drive better than an automatic system. In any event, it seems to be coming and we are all going to have to get used to new ways of doing a lot of things.
Keep in mind that although some AVs will still have steering wheels and controls, others won’t. Google intends to only sell AVs without any controls. So at that stage of the game, you’re really in the hands of the software and hardware.
But I’d still rather take my chances with the extremely low odds of your AV malfunctioning and running a red light and hitting me than with the much greater odds of an inattentive or hurried or intoxicated driver running that red light and hitting me.
Keep in mind that AVs will have powerful self-monitoring systems too. Obviously, they’re not going to be 100% perfect. But if they’re materially less distracted, drunk or feeling macho than humans, than we’re all way ahead.
It’s all about the odds. I’d rather bet against humans than some machines.
Good points. I think the weaknesses of the system are 1) failure potential and 2) whether an almost infinite number of situations can be systematized in order to make the whole thing work as planned.
Programmers have worked for decades to simplify and systematize processes so as to automate them. Here, the programming must accommodate the real world, and not the other way around (as has mostly been the case.) I would feel better if the system were rolled out to work in just a limited number of situations, such as Federal Interstate highways which are engineered and built to particularized specs, thus sharply reducing the variables that must be considered. But out in the big wide world, I don’t have a lot of confidence that everything that can happen will be anticipated and planned for. This might be a small problem or a big problem, but it will be a problem nonetheless.
Accommodating the real world is a problem. I was listening to a Prof form MIT talking about it on the radio the other day (he has been involved with AV for 20 years) he said really humans are surprisingly good when we pay attention, the amount of data we take in and process is astounding when you try to get a computer to do it. Our problem is when we don’t pay attention or are otherwise effected. He did say he knew of no AV that could yet reliably handle the following situations.
Snow,
Hard rain
Hand directed traffic (think construction sites)
unmarked parking lots
Poorly mapped and poorley marked areas (he mentioned several streets in Boston that basically only locals can navigate)
In all of these situations he said in testing Humans faired far better then the computers.
Now all these could be over come but he didn’t think they could be reliably overcome for another 5-10 years.
Also systems do crash believe me even some systems that are never supposed to, do. It’s just a matter of how often, if they are tested and have many backups it should be a minor problem but all that drives up cost.
You also have whats called the trolley problem (now maybe renamed the Mercedes problem) Who does the car save, many people outside the car or the one inside. Mercedes has announced their’s will save the one in the car (their customer) which raises all kinds of ethical questions.
I know it is completely unfashionable to doubt “progress,” but I think we need to be careful as we open Pandora’s box. Electronics and software are not foolproof, even if they are fine 99.9% of the time. If your mobile device has a hiccup (whose doesn’t?) there is no harm done beyond a minor annoyance (unless you are catching fire, courtesy of your Samsung Galaxy Note7). And that’s exactly my point: a reputable, experienced company is suffering through an enormous issue that seemed unthinkable until it started happening. Sure, they will trace the problem, etc., etc., but enormous damage can be done in the meantime.
My current cars–both from the Cyberdyne/Skynet era, are buggy and far less confidence inspiring than vehicles from the same makes (Jeep, BMW) I owned years ago. My Jeep routinely slams on its brakes “automatically” when I am crossing railroad tracks on the main street in my town. The sensors mistake the rising road angle as an obstacle and the Jeep second guesses me and decides to brake. The Jeep has also been recalled to reflash the software to protect it from known hacks (wonder what the unknown hacks will be able to do?). Both the Jeep and BMW have occasional blank screens and reversing sensors/back-up cameras that don’t always respond, etc., etc. Normal glitches I suppose, but since I am still in control, everything is manageable. I know there are plenty of people who will trust their devices implicitly and check-out behind the wheel, and that’s all fine until someone gets killed. The notion that autonomous vehicles will eliminate almost all traffic deaths is nonsense–there is no machine in existence which is that flawless. The machines may be better than distracted drivers, but they are not better than fully engaged, well trained humans.
The human brain remains the most sophisticated piece of “technology” we have, able to decipher and react to incredibly complex variables. Let me provide a real world example that should hit close to home for all readers here. When I was teaching my teenage daughter how to drive a few years ago, she was at the wheel on a residential street where the speed limit is 30 mph. A skateboard came shooting into the street in front of our car, and I yelled at her to stop, which she did. Good thing, because as expected, a kid ran into the street after the skateboard, never bothering to look if anyone was coming. We stopped in time, but I shudder to think what would have happened if my daughter wasn’t paying attention, or the sensors on an autonomous vehicle didn’t detect the skateboard and/or do the instant scenario planning required as to what would happen next. By the time we teach machines to respond and think that effectively, then Stephen Hawking’s dire predictions could be realized.
We stopped in time, but I shudder to think what would have happened if my daughter wasn’t paying attention
Well, that’s about the best argument for AVs there is. The simple reality is that some 35k Americans are getting killed on the road each year, and why? because too many folks don’t pay attention. Distraction is one of biggest reasons by far. And being under the influence.
Look at it this way: How many crashes are caused by mechanical failures? probably very, very few. Almost all crashes (and fatalities) are caused by operator error of some kind.
So yes, if electronic failures cause a tiny handful of crashes, like mechanical failures have been, but the overwhelming majority of human failures are eliminated, than were ahead by a HUGE amount.
Let me put it this way: I trust sophisticated machines a whole lot more than I trust many/most drivers. Sophisticated machines fly our jets (how often do autopilots fail?), monitor our vital stats in the ICU (how often do they fail?), etc…
I always read the articles in our local paper on car crashes, especially fatal/serious injury ones. The causes are inevitably failure to pay attention/control the vehicle. people running across center lines, killing innocents in the process. A mom just went off a road, flipped her SUV, and killed her 4 year old daughter. Failure to negotiate a turn in the road, for whatever reason.
My point is very simple: it’s carnage out there on the roads. 35,000 per year!! I don’t want to get killed or mutilated by another inevitable inattentive or drunk driver. I’ll gladly take my odds on a sophisticated AV over an idiot. It’s the easiest call to make, in my opinion.
The reason why there’s so much skepticism is that everyone (especially here at an auto site) thinks they’re a great driver. In reality, many of us have made mistakes like these that did or could have caused a serious crash. Or came very close to it. Who will throw the first stone? And certainly many of the drivers out there are ticking time bombs.
For your daughter’s sake, I suggest you rethink your position and embrace AVs as a material way to improve her current odds, either behind the wheel or as an innocent victim. No, not by 100%. But wouldn’t you embrace a 90% improvement? Or a 75%? Or even 50%. It’s all in the numbers. Literally.
The human brain remains the most sophisticated piece of “technology” we have, able to decipher and react to incredibly complex variables
The problem is engaging it, and keeping it fully engaged and sober. If one could solve that, we wouldn’t be having this discussion.
No, the skepticism is that this is an auto site and there’s a legitimate skepticism as to what the endgame here is. Is it going to take incentive or inconvenience to ultimately sway the skeptics into this utopian fantasy? Will user operated vehicles be only attainable play things for the wealthy who can afford the penalties(insurance or otherwise).
You used the horse and buggy analogy earlier for automobile skepticism, and, well, as things shook out you can’t store a horse in your back yard or take one down a stroll down the street. Maybe they were right in their concerns. Of course, the difference is infrastructure is completely different for those forms of conveyance, and had to go (as a fan of cars, thank god). But there’s nothing different in infrastructure that will have to change for AVs over the current fleet, so in order to achieve a BIG vehicle fatality drop from the current figure, you’re going to need to run a very good campaign to get the knuckle draggers on board, and if it’s by force, fear mongering and condescension, eh, expect some pushback.
As for 35k vehicle fatalities, deaths from obesity trumps that seven fold. So now we’re going to free up people’s hands completely? Let’s just say I’m buying stock in fast food.
No, the skepticism is that this is an auto site and there’s a legitimate skepticism as to what the endgame here is.
No one has that answer. AV will slowly be implemented over decades. And they will be highly optional. I sense a lot of anxiety about being forced into one. I don’t see that, unless you’re proven yourself to be irresponsible behind the wheel. There may be incentives, like lower insurance or other ones.
There are many situations and locales where AVs will not work, or at least for a long time. In rural places, or where off-road use is involved, or other situations, it’s juts not feasible.
In any case, it’s impossible to predict how AVs will integrate into our lives. Clearly, they are attractive to certain groups; a number of engineers working on AVs don’t even drive or don’t like to! That’s the point I’m making about this being a car site. There are many people out there for whom driving is a chore and a waste of time. They’d much rather be doing something else.
So AVs will firts show up in certain metro areas, as taxis perhaps and for those that chose to have the convenience. To tell you the truth, there’s times I’d like to have it right now. I run too many needless errands.
But worrying about having your fingers pried off your beloved steering wheel are grossly premature. I know I have no intentions of having that happen to me. But I’d like the option.
a number of engineers working on AVs don’t even drive or don’t like to!
Duh lol
I do hope you’re right, my main concern is that incentive doesn’t transform into penalty, where “lower insurance” isn’t just “same insurance”, but lower than what non-AVs will now be paying
I mostly doubt driving will be taken as anything less of a chore if the car does it for you however. Sometimes you may be able to accomplish work, tasks or entertain yourself in ways you couldn’t do otherwise, but if you’re just sitting there twiddling your thumbs like in this video, I feel like all the stresses of driving are still there, seeing other cars pass you, being in the “wrong” lane and all the other sensory stuff you pick up from manually driving, will still stress you out. I get stressed out in those ways riding as a passenger with someone else at the wheel, even if I’m occupied and don’t want to drive at that moment. Maybe that will be different when an entire generation is exclusively raised on AVs but it’s still going to be a waste of time and a chore(if only in payments and maintenance).
but if you’re just sitting there twiddling your thumbs like in this video, I feel like all the stresses of driving are still there,
That’s only because it’s legally required under the current regulations. With L4-5 full autonomy, no need to pay any attention at all. Turn the seats 180 degrees and watch the big screen in the back of the car. That’s really the whole point, and the benefits of it, for the driver.
Your points highlight why this issue is so critical for us to address thoughtfully. I agree that human error (much of it distracted and/or impaired driving) accounts for a tremendous number of accidents/fatalities, but I’m not sure the answer is to cede control to the machines entirely. Software engineers are humans too, and they can make mistakes. Hackers are human, and they can mean to cause harm. Humankind never does well when hubris exceeds reality: witness the “unsinkable” Titanic.
Enormous progress has been made through the years in terms of automotive safety. Traffic fatalities have dropped 20% in the last 10 years and the per capita death rates are at (or near) an all time low. Vehicles protect occupants better than ever, and this has led to a significant reduction in fatalities, even though distractions, in the form of mobile devices, has increased markedly. But we should not allow people to take the easy way out and just get lazy and complacent and not take responsibility for their own driving. Yet that is exactly what we are doing. The scary reality is that there will inevitably be a long “hybrid” period where autonomous and non autonomous drivers share the road. If everyone feels that they can just check out, I guarantee that mayhem will ensue.
Perhaps we should make mobile device makers, carriers and app developers liable for injuries and deaths while consumers are distracted (driving or not)–that would open up an enormous pool of money for trial lawyers and would change behavior (for both consumers and tech companies) quite rapidly. Imagine if your car and your phone communicated so that you would be blocked from texting and driving. That alone could yield a large percentage improvement in current accident statistics. Or what if we initiated legitimately rigorous driver training, like what is commonplace in Europe? Some tech innovations are already helping: I’d argue that Uber is getting intoxicated drivers off the road by being the readily available “designated driver.” Common sense technology can augment, but not replace, human decision making. But people should never be lulled into giving up their responsibilities entirely.
I am a huge technology advocate and I willingly embrace change, but I think we need to be realistic about timelines and ways to implement the technologies so that they are genuine improvements and not just rapidly implemented utopian ideals that may prove suboptimal after real world user testing. For our daughter, my wife and I have felt strongly that great training is her best defense. We had her complete a rigorous driving school program and mobile device use in the car is strictly forbidden (my wife and I set that example: we do not engage with our devices while driving). Training her to be an aware and defensive driver is still her best protection for the foreseeable future, rather than just turning herself over to an a/v that may or may not be fully capable to protect her in every situation (whoops, we missed that bug, sorry… but we promise it will be fixed for the next software release). Plus, she really doesn’t need to be online 24/7–she can take a break in the car for the duration of a journey. God forbid she actually looked at the world around her rather than just a screen.
Every time anyone gets behind the wheel there is risk. You do everything possible to mitigate that risk, but sometimes luck is not on your side. My Pop was an excellent driver. He was also one of the 40,150 traffic fatalities in 1993 when he was hit by a drunk driver. Whether it was an errant robot or an inebriated human that was responsible, the tragedy happened nonetheless. Machines prolonged our family’s agony, as Pop was rushed to the hospital in a coma and put on life support. He had no brain activity and was pronounced brain dead, but he was considered “alive” because he had already been hooked up to life support and we had to initiate the process of triggering his living will (weeks and weeks of wrangling even though his paperwork was in order and the prognosis for any sort of recovery was zero). Pop was certainly not alive at that point, though the machines could keep his lifeless body going in a pitifully diminished state. Mercifully Pop passed before any plugs had to be pulled–we like to think that his spirit had the gumption to say “wtf is this s**t?!?! I’m outta here!” Human until the very end.
We should always ask ourselves: just because we can do something, does that mean we should? Have we considered the unintended consequences of our actions? Have we overestimated our own abilities to deliver an “invincible” solution to our problems? No easy answers, without a doubt.
A number of good points, many that can’t really be answered. But I’ll adress a couple.
Imagine if your car and your phone communicated so that you would be blocked from texting and driving.
It’s readily available; phones could be cut off by use from drivers, but phone makers won’t implement it:http://www.nytimes.com/2016/09/25/technology/phone-makers-could-cut-off-drivers-so-why-dont-they.html?smid=tw-share&_r=1
Or what if we initiated legitimately rigorous driver training, like what is commonplace in Europe?
The faith in this is misplaced. The simple fact is that almost all crashes are caused by the most basic and simple things, like distraction, poor judgement, road rage, inebriation, etc. They are hardly ever caused by the lack of higher technical skills or knowledge. There is simply no evidence that supports that more drivers ed improves outcomes. The most critical skills are attention and reasonably basic good judgement. Anyone can bring that, if they chose. One can learn to drive in 10-30 minutes, but the decision as to whether to do it with good judgement and attention can’t be taught.
Software engineers are humans too, and they can make mistakes.
The question is if they make more mistakes in AV software than drivers do behind the wheel.
In reality, these software systems are of course very complex, and the validation process will involve millions of miles. And of course these are essentially learning systems, a form of artificial intelligence. As more information is accumulated from cars on the streets, the better they will perform. But will there be glitches? Aren’t there always? The question is how serious they are.
We should always ask ourselves: just because we can do something, does that mean we should?
35k killed in car crashes per year equals one full 747 going down per week. If that was happening, would we do something, including something that might involve some risk?
This is not being done just for kicks or the money only; it really has huge potential to reduce deaths. Given the current costs, some inevitable glitches along the way are undoubtedly well worth it.
Of course it shouldn’t be rushed or done cavalierly. And there is a risk in being too far ahead of the technology, and Tesla is at risk of that here. But if we can reasonably expect deaths to be reduced by a significant extent, then yes, we absolutely should do it. With care and diligence.
To not do it would be highly irresponsible, given the current grim reality.
Call me a Luddite but I think self-driving cars are a terrible idea. Can’t wait to see how they’ll handle the unpredictable winter weather up here in Canada. Hummph!
That may take a while. It will be a good challenge for them. There are some locales an circumstances where AVs may not work for some time.
My CTS front sensors don’t like the poles that divide the lanes when all the traffic is routed off one set of traffic lanes during road repair. It periodically flashes an impending crash in the heads up display and buzzes my seat, but before slamming on the brakes it seems to know it was a false alarm.
There is also the issue of hackability. I can guarantee that cars that have wireless communication systems will be hacked. You can also be certain that for every security update there will be a new hack.
http://thenextweb.com/gadgets/2016/09/21/tesla-races-to-issue-security-update-after-researchers-hack-its-model-s-from-12-miles-away/
I’ll be sticking with my old-school, non-connected car(s) for the foreseeable future.
QUOTE: I’ll be sticking with my old-school, non-connected car(s) for the foreseeable future.
Ditto here. They will have to pry my cold dead fingers off the steering wheel before I give up the pleasure of driving my own car. Nobody has the right to tell me my car will drive me by itself and I have no choice in the matter.
It’s a personal capsule. When the technology is ubiquitous, people will look back in horror at that fact that we we actually would put ourselves in such danger on a daily basis, rather like the time before machines had guards and automatic safety cutoffs.
“Yes, kids. There was a time when everyone controlled their own car.”
“Daddy, even Mr. Hobson?”
“Yes, even Mr. Hobson. People like him would be on the road, controlling their own cars all the time, until their kids took their keys away… or they had and accident and someone was injured or killed.”
“I’m scared, Daddy!”
“Don’t worry, Honey… those days are long gone.”
How long will it last and how much will it cost to repair when it fails? As it is, late model cars are subject to failures of their complex electronic and mechanical systems that make them uneconomical to repair once the warranty runs out.
In the future will one be able to pick up a 15-20 year-old autonomous car for cheap and reasonably expect to keep it on the road with with a little wrenching and TLC?
Most mechanics I have spoke to say that cars/SUV’s within the past 5-8 years are getting to be a nightmare to not only keep up and inspected but are also getting more and more difficult to diagnose correctly. I hear the same complaint from consumers over and over again from the dealers when bringing these vehicles in for warranty repairs. “Cannot replicate or duplicate problem/concern”. And year after year the complexity keeps going up so I would expect more of the same in the years to come.
OK, I get it now…
The ancient piece of wood that I have in my hand drives the point home. I am holding a beautiful Keuffel & Esser slide rule. It is the same one which got me through high school and college. I have not used it in 35 years, and I am embarrassed when I realize that I don’t remember how to work the darn thing. I can’t even find the on/off switch 🙂 The point is the ability to use a slip stick is obsolete. I replaced it with a keypad and never looked back. The superiority of the microprocessor cannot be argued.
So here we are at the eve of the revolution where manual operation of an automobile will become as obsolete as my slide rule. In time, universal adoption of this technology is a foregone conclusion. It is just a matter of time. Hopefully this new technology will prove equally superior beyond question.
I hope that we are wise in our aspirations, but one thing nags on my mind. – We have all heard the sentiment that smarter cars breed dumber drivers for a long time. Five minutes on any road anywhere will confirm the fact that cars have gotten pretty smart.
A car that drives itself, rendering the human driver unnecessary, would appear the pinnacle of this trajectory.
Where does that leave us? We will have successfully built a system that is certainly safer and more efficient, but are we really better off in the big picture?
All arguments against Automatic Drive are irrelevant if it knocks the death rate down significantly. Who wouldn’t sign off on a technology that cuts deaths down from 35,000 to 3,500… or even 350? People would still be killed by the ones who don’t have it, but those who do would have max chance of avoiding that fate from without. I was very doubtful on it before this discussion started, not wanting to give up control, but I have completely reversed my position. Giving up the wheel just doesn’t mean giving up control. It’s still you and your car, the same team. You haven’t been forced to take mass transit. You can program it to drive as fast as it can safely go, which is faster in some cars than in others. You want thrills? It’s like being on a roller coaster. On another day, you might want to look at the leaves, so you put it in Sunday Drive mode. If you must race, you don’t do it on a public road because your car won’t let you. Even if some “bad hombre” is driving recklessly without Auto Drive, your car can react faster than you can to avoid him.
The point is, you are still in control, but now you have so much less to worry about.
All of this can be done by voice recognition. Change your mind suddenly? The car reverses direction, but it doesn’t put you or others in danger doing it.
How about road rage? Where’s the excuse for that anymore? The car won’t let you act out. Even if someone is pissed that his car won’t make an unsafe pass, the car doesn’t care. It keeps him safe whether or not he deserves it.
An odd thing happened to me on my way home from work. Driving on two lane roads with this discussion in my head, I nearly panicked to think that the 3500 lib vehicles separated from me by just a painted line, and traveling at a closing speed of between 6o and 100 miles per hour were under the control of human beings.
There are plenty of places where we want and should have total control. We aren’t good enough to make that argument for ourselves as drivers.
Not that I’m advocating this activity, but theoretically an AV street race should be super safe, right? I mean, they’re supposed to have sense of their surroundings, they’re supposed to be aware of other cars, and are supposed to outperform us mushy scary human beings, I cannot possibly see what can go wrong.
I’d trust my decidedly low-tech ’64 Falcon before I’d trust ~any~ computer-driven/self-driving car. Just the way I am. Give the folks that wanna get around in a self-driving car their own lane(s) and keep those poxy high-tech ‘things’ the hell away from me.
as a complete offside paul, I admit I didn’t check out the podcast but am wondering if the alex roy mentioned with your son is the alex roy of cannonball fame?
if so your son runs in excellent company! he is a real character!
AVs raise a difficult question of ethics many scientists and legal eggheads are working on.
Basically, the question is how do you apply Isaac Asimov’s robotic laws to AVs.
In particular, if the AV gets into a position if it has to choose between two evils, meaning killing the driver or a pedestrian, who should it choose ?
In other words, should the AV must be programmed to choose its driver, or the pedestrian ?
Should it choose its driver and passengers over ONE pedestrian ? Or the contrary ? Should it choose its driver and its passengers over MANY pedestrians ? Or the contrary ?
If the car company chooses to or is forced by the regulator to choose the car occupants over pedestrians, will people accept to buy and/or run a car they know to be programmed to sacrifice them over pedestrians ?
I think today’s cyberattack that has impacted large swaths of the internet is interesting to consider as part of this thread. According to Bloomberg, the hackers have been able to co-opt the poorly secured “internet of things” to wreak havoc:
“The barrage likely originated with a large amount of poorly secured devices like internet-connected cameras, routers, and digital video recorders, according to an analysis of the attack on Krebs’s site. These devices, collectively referred to as the “Internet of Things,” have been the source of an increasing number of attacks since early 2015, Flashpoint and Level 3 Threat Research Labs said in a report published last month.”
Here is the link to the full article: https://www.bloomberg.com/news/articles/2016-10-21/internet-service-disrupted-in-large-parts-of-eastern-u-s
While it is not the end of the world to not have access to Twitter and Spotify for a few hours, the results could be far more severe if a/v operation could be compromised–you can bet there will be malicious hackers seeking to do just that.
Other thoughts. Driven cars presumably won’t go away overnight in favor of AVs, in fact I’d say the same forward thinking technologists highly in favor of this will tell you just how advanced and reliable the current fleet is, and has been for some time now. In the event that AVs were to be granted privilege on roadways, or even lead to the full ban of driven vehicles, what kind of effect is that going to have on the used market?
The supposed NTHSA goal being zero deaths within 30 years as mentioned above is beyond unrealistic unless you completely screw over swaths of the population who can only afford driven cars on the used market, think about it, there’s perfectly good 20 year old cars driving around today in 2016, and technology on those weren’t even as advanced as they are today, so one can only presume the 2016s will be around as long or even longer, so all of these cars, even the recently used ones are rendered unviable? Gonna have some very pissed off people when their resale plummets greater than a 1966 studebaker.
Cars aren’t like steam engines where technology can simply be usurped by the likes of diesel electric and render the old way obsolete for service, there were a handful of railroads and a couple thousand locomotives in that transition, not tens of millions of private car owners. Hell not everyone owned a horse and buggy during the dawn of the automobile either, so even that analogy isn’t apt for what this transition would be like.
It’s very interesting in a disturbing way to see how this whole notion of a AVs has evolved in such a short time from what is essentially an advanced cruise control like feature into this faux savior of all mankind where resistance or cynicism towards it is deemed futile, suicidal and even and selfish. I mean some of the fear mongering rhetoric thrown about in this thread is akin to TRUTH ads. “This is your brain, this is your brain when you drive” *smash*
I predict that where AV cars predominate (urban areas) that you won’t want to drive anyway. A bunch of vehicles going precisely 25 mph (max) in synchronicity will just lock you into a pattern best handled by another AV.
I also predict that said urban areas will be clogged with cars because it will be cheaper to send your AV on a sight-seeing tour without you than to park it at a meter or garage. .
If the tech won’t let the car travel unoccupied, then this might do the trick:
https://xkcd.com/1559/
But there’s no reason AV’s shouldn’t travel unoccupied. As long as AV’s sell, gasoline is being burned/electricity used, tires etc are being consumed, and someone is paying for it– then the goals are attained.