Tesla’s Elon Musk: Full self-driving technology nearly complete
We have self-driving transportation. It’s called high-speed rail augmented by local trolley systems. The privately owned automobile is obsolete, and so are its advocates. Now they don’t even want to drive the damned things themselves. They want IT to drive. Stop the absurdity. Ditch automobiles for the good of humankind’s future.
I take it you don’t live in 99.5% of the United States, or probably 99.9% of the world?
Self-driving trucks already rolling on Florida highways
Decades, I tell ya…
Highway driving is easy.Full autonomy is not.
It seems to me that providing just the highway driving is enough to reach a huge market. And is realistically do-able soon.
But the car companies all seem to be competing for the glory of the ultimate prize (full autonomy), even if it takes years and billions of dollars.
They could win a lot of sales in the meantime, by selling a partial solution.
But nobody seems to be even trying. Why not?
Lots of people would love be able to sleep for a couple hours when they drive on longer trips.
There must be millions of people who don’t visit their grandkids often, because it’s a 4 hour drive each way.
Or business people who’d like to personally supervise the branch office in another city once a week.
These things may be too much hassle for regular driving, but if you could sleep most of the way, it becomes easy.
Why isn’t anybody aiming for this market?
As for Elon Musk’s declaration that “next year” his car will work perfectly: I won’t believe it till he actually produces something and uses it in the real world.
Is this the same Elon Musk who promised a Falcon Heavy launch in 2013 or early 2014?
Timeline of Tesla developments 2017-2018
Because while cruisng on the highway takes relatively little skill, driving in backed up traffic with drivers erratically chang8ng lanes is just as difficult as navigating urban streets. And we’ve already seen systems like Telsa’s self-named “Autopilot” fail dramatically even in completely benign conditions due to not being able to recognize an obvious hazard.
It isn’t sufficient for a truly autonomous piloting system to be capable of dealing with 95% of driving scenarios, or even 99%; it needs to be capable of handling any and all situations a competent, alert human driver can handle with a high degree of reliability, else no reputable vehicle manufacturer will field it (or like Tesla, will try to hide behind indemnification statements that their so-called “Autopilot” is just a driver assist system). Over-the-road bulk cargo transportation is actually a more plauible entry into autonomous vehicles because a national carrier could establish regional distribution hubs near highway exits and avoid most traffic issues, lessening the problems of hiring and retaining long haul truckers and the limitations that come with them, as well as implementing battery electric vehicles which can carry large battery packs with the range for many hundreds of miles and cruise efficiently at highway speeds.
Autonomously piloted passenger vehicles still have a way to go before general adoption, and will very likely have a different sales and service model than passenger vehicles do today. When they do achieve a threshold of reliability in traffic that is better than human drivers (and they will given the all-around situational awareness and lack of fatigue or distraction) the costs of insurance alone will drive wide adoption, notwithstanding convenience, but claims that this is just around the corner of next year or two are hopelessly optimistic.
Stranger
Good argument. The standard of living in the United States, and our technology and infrastructure, should generally reflect the standard of living commonly enjoyed by the rest of the world.
You’re really on to something here!
I think long-haul trips on limited access highways may be achievable pretty soon but the problems of town or city driving will be around for a long time – too many things in states of constant change, too many unpredictable movements by other cars, bicycles, pedestrians, waving flags and windblown bushes and tree branches, sidewalk cafes, etc., and the teenage boys making a game of faking stepping in front of the autocars to get them to swerve or slam on the brakes.
Another big problem I recently read about is the parking problem – self-driving cars will not want to pay the very expensive parking fees in cities; they will find it less expensive to simply keep driving around until they are called for a ride – thousands of cars endlessly aimlessly circling the block at very low speed (because it cheaper to go slower), jamming up traffic.
I live in a major metropolis with four seasons where roads are opened without painted lanes.
Snow clearing and removal has been terrible this winter, such that drivers are frequently stuck in ruts – and get bounced out because of a stray lump of ice.
I have serious doubts any autonomous car will be able to operate in these conditions in the next decade.
We’ve already had self-driving cars on local roads, with a person behind the wheel ready to take over at a moment’s notice.
Now you’re pointing out that the same thing is happening on highways. And highways are generally considered to be a simpler situation to deal with than local roads are.
So if I understand this correctly, they’ve made the same progress in solving an easier problem as other outfits managed years ago in solving a harder problem.
IOW, so what?
This isn’t an argument against “self-driving” passenger cars; it’s against the private ownership model of such vehicles. Autonomously piloted vehicles operating in a subscription rideshare fleet model could minimze congestion and vastly reduce the necessity of expensive parking in city centers and the wasted real estate of gigantic business parking lots. By serving multiple functions and increasing utilization—commuting during rush hours, shuttling children to after-school activities, and package/grocery delivery during low usage hours—the subscription costs could be reduced well below the cost of owning and maintaining a car, allowing households to get by with just one or no privately owned vehicles.
Part of the problem with how people project how autonomous vehicles will be used is that most people imagine the future to be just like today except robots chauffeuring them around in their own vehicle, which is like imagining interstate highways populated with Model T cars from the 1920s. In fact, autonomously piloted vehicles would have the same kind of social revolution as the washing machine or the smartphone, opening up inexpensive and readily available transportation to anyone in an area populated enough to justify fleet access. Combined with reduction in accidents (and hiher insurance premiums for human drivers who insist on operating their own vehicles) it will almost certainly cause private vehicle ownership to be a luxury if not a niche market while future generations don’t even bother learning to drive any more than they bother mastering the skill of churning milk for cream and butter.
Stranger
We already have subscription services and shared vehicles.
They don’t minimize the number of vehicles on the road or private ownership.
Instead, they become an interim step between not owning and owning a private vehicle, or a cheaper option instead of owning a second or third vehicle.
That’s not my opinion – that’s how companies like Car2Go are marketed.
And when it comes to ownership people do make the fiscal choice based on how much they drive and they balance the cost of ownership vs. subscription. A co-worker of mine who was driving a shared car a lot over the past three years bit the bullet and bought a vehicle last year because her monthly costs were getting in line with private ownership.
Public fleet (“shared vehicle”) programs are primarily to people in high population density areas, e.g. people who live in or around city centers, and are intended for use by people who do not need daily commuter use of a vehicle. By themselves they don’t minimize congestion because it is a niche market, and as a niche market, fleet operators will charge what the market can bear, e.g. costing users just marginally less than the total costs of ownership or for people who do not have a place to storage a vehicle. They also don’t off the convenience of being able to transport while not having to drive, or sending the vehicle to pick up groceries, packages, or kids without having to go with it.
The point here is that there is that with autonomously piloted vehicles, combined with battery-powered electric vehicles which can fungibly change their power source to whatever the cheapest form of electricity generation is and don’t have the emissions issues of gasoline or diesel powered vehicles, the degree to which they can be shared and otherwise utilized is no longer dependent upon individual owners or the small population of people in a dense urban area subscribing to the surface. Any area with a population density to share vehicles between a significant pool of users–and with autonomous piloting, those users can include below-driving-age adolescents as well as non-passenger users–can be profitable to a fleet operator in the way a Zipcar or car2go cannot.
And if autonomous vehicles become more reliable and less prone to causing accidents–which they should be capable of doing once they achieve a degree of autonomy equivalent to that of an alert human driver–the cost of insurance (as the overall number of accidents go down and insurance companies have to lower costs to autonomous fleet operators to be competitive, which means raising premiums on the more accident-prone human drivers) will likely drive many would-be owners away from possessing multiple vehicles for commuting (although it is safe to assume that many people may still own vehicles for recreational purposes or because they have particular needs not met by fleet vehicles such as offroad use). It is notable that the possession of a drivers license is declining not only among millennials but across overall demographics as people tend to migrate toward urban environments for vocational and recreational opportunities. This will be exacerbated by the aging baby boomer generations who are living longer and getting to the point of not being able to safely operate a vehicle but still wanting the freedom of being able to go places without relying on public transportation or rides from family or others.
Currently many of these people rely on rideshare companies like Uber and Lyft, but these companies themselves understand that employing (or more properly, contracting with) human drivers is an interim business practice that will come to an end once Level 5 autonomous vehicles become a reality, because the degree of utilization and the elimination of the often problematic human element will make the use of human drivers too costly. The shift to autonomously piloted subscription-based fleet services will be due to the confluence of different technologies combined with socioeconomic changes that are already occurring, which will make private ownership of a vehicle more of a luxury, and an expensive one at that.
Stranger
Why do you think it will be particularly expensive? I can see what you’re saying about the price of Ubering everywhere plummeting, and certainly if that were a reasonable option a lot of people would opt not to own a car with all of the costs that go along with it (parking, insurance, maintenance, etc). But autonomous cars aren’t going to be more expensive than cars are right now, and there’s lots of people for whom these costs aren’t that big of a deal. I currently have 3 cars, all bought used, all non-luxury marquees, and the costs associated with them aren’t really hurting me in any way. I’m sure my setup would be considered luxurious to some but it doesn’t feel that way, and I don’t see myself giving up private ownership even if the cars were autonomous.
I know, right? This has become such a repetitive conversation.
“An autonomous road vehicle has just been launched!”
“Oh, cool! So nobody behind the wheel or remote monitoring to take over?”
“Well, yes, there is…but still, it’s nearly here!”
Realistically, over the next decade if automakers focus on level 4 vehicles (like the new Tesla plant does) and ADAS (which will provide a huge social benefit in accident reduction) there will be plenty of progress. But level 5? Many decades away.
I see those interpretations of fewer drivers and think they are missing a key factor: the shift in spending patterns and job security.
Until 20 years ago if you wanted freedom and access to people you needed to be physically present.
Today? It’s more important to be virtually connected, and so people are spending large sums of money on computers, phones, and internet connections. Money that a generation ago was spent on physical mobility and savings.
As a result many people are just delaying acquiring a driver’s licence, (or moving out, or getting married, or having kids) but finally going through the steps their parents did, just at an older age.
My wife and I happen to have lots of friends who are twenty years younger than us.
Those with good paying jobs (ie doctors) are buying cars and getting married, while those who are not as financially stable are staying with parents.
Well, according to the article it will be a “few years”. 10 sounds about right.
This article has an expanded quote from the same podcast mentioned in my earlier post:
Elon Musk: Tesla drivers can sleep behind the wheel ‘next year’
How much longer will people need to keep their hands on the wheel while driving a Tesla?
“At least six months from here” -Elon Musk, April 2019.
He’s not wrong.