Self-Driving Cars Hit the Wall

A new article in the Guardian by Christian Wolmar, author of Driverless Cars: On a Road to Nowhere, summarizes how the grand promises for the self-driving vehicles have fallen well short as lawsuits mount, cities impose restrictions, and the UK has gotten tough on overselling driving-assistance. Wolmar depicts the autonomous driving vehicle industry (save ever-stubborn Tesla) increasingly in reverse gear, with one time leader wannabe Uber having exited the business and GM (via its Cruise operation) floundering.

Wolmar provided a fine compact account, but one is still left wondering, how did such a not-ready-for-prime-time technology get so far? Much of the impetus resulted from the Silicon Valley attitude of “Move quickly and break things,” as demonstrated particularly with the deliberately mislabeled ride sharing companies Uber and Lyft. They rolled over local cab regulations and got little resistance, largely because taxis are local businesses, with not a lot of capital, and no meaningful constituency. Hubert Horan has chronicled in exhausting detail that the gig taxis never made any sense, that they had inherently higher costs than traditional cabs.

But self-driving cars, as in self-driving taxis, were a big part of the Uber/Lyft hype. Just imagine how great these companies’ economics would be if they could be rid of the expense of drivers! Of course, that argument ignored that those seen-as-disposable drivers provided the car, its maintenance, its insurance, and its gas, and typically did not understand their economics (as in saw their revenues minus gas as their profits), all the better for the ride-hail companies to exploit them. Uber and Lyft with driverless cars would mean Uber and Lyft having to invest in and own fleets. How could that big capital outlay and additional overhead possibly improve their economics?

We were skeptical of the idea that it would be possible to arrive at autonomous self-driving cars, and our concerns were confirmed by the fact all these supposedly self-driving vehicles had human oversight, as in either a driver in the car or remote human oversight. We thought that had the potential to increase hazards, since a non-driving monitor would easily space out, and then be forced to snap back to attention when the car sent an alert, and would be in cognitive catch-up mode as to what was happening.

To put it another way, the only way self-driving cars appeared able to live up to their promise would be if all cars were self-driving and ceded control to a central network, so the network would control the behavior of all vehicles, greatly limiting hazards and random events. But the “Level 5 autonomous vehicle” was based on the premise that somehow individual cars would become so smart and so good at data-crunching that they could navigate successfully even with many vehicles all operating independently, and many with pesky drivers. For instance, from Car Magazine:

The difference between Level 4 and 5 is simple: the last step towards full automation doesn’t require the car to be in the so-called ‘operational design domain’. Rather than working in a carefully managed (usually urban) environment with lots of dedicated lane markings or infrastructure, it’ll be able to self-drive anywhere. How? Because the frequency and volume of data, the rapid development of artificial intelligence (AI) and the sophistication of the computers crunching it, will mean the cars are sentient. It’s a brave new world – and one that Google’s Waymo car is gunning for, leapfrogging traditional manufacturers’ efforts. The disruption will be huge: analysts HIS forecast 21 million autonomous vehicles globally by 2035.

Wolmar describes how this Brave New World has stalled out. The big reason is that the world is too complicated. or to put it in Taleb-like terms, there are way too many tail events to get them into training sets for AI in cars to learn about them. The other issue, which Wolmar does not make explicit, is that the public does not appear willing to accept the sort of slip-shod tech standards of buggy consumer software. The airline industry, which is very heavy regulated, has an impeccable safety record, and citizens appear to expect something closer to that…particularly citizens who don’t own or have investments in self-driving cars and never consented to their risks. From Wolmar:

Developing driverless cars has been AI’s greatest test. Today we can say it has failed miserably…. Moreover, the recent withdrawal from the market of a leading provider of robotaxis in the US, coupled with the introduction of strict legislation in the UK, suggests that the developers’ hopes of monetising the concept are even more remote than before. The very future of the idea hangs in the balance….

Right from the start, the hype far outpaced the technological advances. In 2010, at the Shanghai Expo, General Motors had produced a video showing a driverless car taking a pregnant woman to hospital at breakneck speed and, as the commentary assured the viewers, safely….

First to go was Uber after an accident in which one of its self-driving cars killed Elaine Herzberg in Phoenix, Arizona. The car was in autonomous mode, and its “operator” was accused of watching a TV show, meaning they did not notice when the car hit Herzberg, who had confused its computers by stepping on to the highway pushing a bike carrying bags on its handlebars. Fatally, the computer could not interpret this confusing array of objects….

Now Cruise, the company bought by General Motors to spearhead its development of autonomous vehicles, is retreating almost as rapidly…In October, a woman crossing a road in San Francisco was hit by a human-driven car and knocked into the path of a Cruise robotaxi. Instead of stopping, the robotaxi drove over the pedestrian because it had been programmed to pull over to the right when confronted with an unknown situation. She survived but will clearly be in line for massive compensation…

Cruise… soon withdrew its robotaxis in all US cities and its CEO quit. It was revealed that vehicles were not even driverless, since the cars had been remotely controlled with interventions by operators about every four or five miles. There are now mass redundancies and the future of the development is uncertain.

And the UK has Tesla in its crosshairs:

In the US, where there have been numerous accidents with Teslas in “full self-driving” mode, the manufacturer is facing several lawsuits.

In the UK, Tesla will fall foul of the legislation introduced into parliament last month, which prevents companies from misleading the public about the capability of their vehicles. Tesla’s troubles have been compounded by the revelations from ex-employee Lukasz Krupski who claims the self-drive capabilities of Teslas pose a risk to the public. Manufacturers will be forced to specify precisely which functions of the car – steering, brakes, acceleration – have been automated. Tesla will have to change its marketing approach in order to comply. So, while the bill has been promoted as enabling the more rapid introduction of driverless cars, meeting its restrictive terms may prove to be an insuperable obstacle for their developers.

Ironically, this industry looks to have created its own woes by going off road early on. The initial impetus was a DARPA Grand Challenge in the early 2000s to build a vehicle that could drive long distances across the desert autonomously. Profit-made visionaries quickly expanded the use case to driving in built environments with their much greater complexity and risks, particularly to other people.

Perhaps these promoters should remember the lesson of the microwave. Appliance companies spend decades promoting the microwave as a replacement for the general-purpose oven, when it is lousy at lots of cooking tasks, like baking and browning. It was only when designers realized that its proper use was as a limited-use rapid heating device that sales took off.

Print Friendly, PDF & Email


    1. pay the piper

      or autonomous parking, assisted cruise control, adaptive acceleration, and adaptive braking… and then we stop. I was one who was very excited for the prospect of self-driving tech, but have come to realize the tech does not have to remove humans entirely, just enough so that driving is safer.

    2. Yves Smith Post author

      I am a demon parallel parker, way better than angle or head in parking. Can do it in tight spots, one pass. It’s my idiot savant driving skill. Would be sad not be able to show it off.

      1. FreeMarketApologist

        Yes! I learned to parallel park a big ’70s era vehicle that short little me couldn’t see the corners of. Got very good at parking (thanks Mom for spending the time practicing with me), and am not giving up my ability to show it off.

  1. SocalJimObjects

    I read the article and I am curious about the lack of mention of Google’s Waymo. Per Wikipedia, “Waymo currently operates commercial robotaxi services in Phoenix, Arizona and San Francisco, with new services planned in Los Angeles and Austin, Texas.” Public criticism on Cruise was fully deserved, but I have not heard the same level of antagonism directed at Waymo’s self driving cars or perhaps the media has been selective in their coverage.

    As someone who hates driving, a self driving car would be a godsend, but I guess we can’t have nice things. Another alternative to a centralized system is to massively simplify the driving environment but if that were to happen, a self driving car won’t be required in the first place.

    1. Carolinian

      I believe Waymo in Phoenix is still confined to one neighborhood and is not city wide. And San Francisco has just passed new restrictions aimed at Cruise which has withdrawn after various incidents.

      “It was only when designers realized that its proper use” says the above about the microwave and arguably self drive may yet live if confined to controlled access freeways as opposed to the far more chaotic urban environment. It’s not that self drive is such a terrible idea given the many problems with human drive and our ever more crowded highways. Undoubtedly as Yves says the hubris is about AI and the notion that machines are now as smart as people which they aren’t.

    2. albrt

      Waymo is what we have here in Phoenix AZ, so I would also be interested to hear about them. They’re very common. I’m extremely skeptical about driverless cars, but I have to admit that from my perspective as a bicyclist they seem to drive more carefully than most humans.

      Waymo spent years getting these cars acclimated with a backup driver in them and no passengers. They were so ubiquitous for so long that people were becoming paranoid that it was just a scheme to spy on us and the cars would never actually go driverless.

      The learning curve was probably made easier because of the lower traffic during the pandemic. I doubt they can ever make back the money spent on having the cars learn the streets, and I have no idea whether being able to navigate downtown Phoenix makes it easier to learn a new area. I also don’t know whether they have remote human backups.

      1. Carolinian

        Thanks for the update. Of course much of Greater Phoenix–with those wide boulevards–is nothing like San Francisco.

        1. albrt

          And it’s not greater Phoenix, it’s just a small part of the city – Lesser Phoenix I guess.

          The Google cars also look wildly expensive – they’re specially made Jaguar SUVs with multiple kinds of sensors sticking out at all angles.

          1. ksw

            Waymo offers service over 225 miles in parts of Phoenix, Scottsdale, Paradise Valley, Tempe, Mesa, Chandler see map:,Scottsdale%2C%20to%20the%20East%20Valley. They are common here in the central core of the metro. So far, I haven’t seen them making bad moves. I looked into seeing if they go to Sky Harbor, the main airport. They drop off at a sky train terminal so aren’t actually going into the main drop off areas. Phoenix metro has second highest rate of pedestrian deaths in the country and Arizona has the most aggressive bad mannered drivers in the country so begrudgingly, I’m watchful waiting on Waymo. Waymo cars aren’t drag racing nearby or taking over intersections doing burnouts.

    3. lyman alpha blob

      About 100 years ago, we had an extensive network of electric vehicles that could take people who didn’t like to drive from place to place fairly rapidly. They were called trolleys, and lots of large cities had them. Too bad we demolished them all rather than expanding them in order to make automakers rich.

      1. Mike Smitka

        Robotaxis never made sense to me, taxi drivers do much more than just drive – helping with luggage, cleaning the vehicle, handling riders with walkers – and the commercial case requires high ride density, which means only the denser and richer parts of cities, given that the requisite tech makes them an expensive piece of equipment. Good luck finding a robotaxi in bad weather on Friday evening in Manhattan, just as with human-driven taxis there can never be enough for when you really need one.

        And I live in a rural area, there was one taxi for 35,000 people, but it couldn’t make ends meet, there’s now only a non-profit service that focuses on scheduled rides for those unable to drive, and on weekdays. And of course, asking neighbors for help.

        See Hubert Horan’s many excellent articles on NC!

        BUT … but passenger cars are not the only use case for autonomous vehicles. Inside factories, in retirement communities, in logistics centers, for trucking on regular routes on highways, for fixed-route, non-highway-speed urban shuttles, for local low-speed package delivery robots, and for agriculture and mining – all niches, but with actual business cases and lower technical hurdles. In the background cell phone networks in many countries are now sufficiently robust to allow remote monitoring and even, where appropriate, brief periods of remote operation where the “robot” isn’t expected to work or when it signals problems. It isn’t necessary to “solve” 100% of the driving problems, as long as one “supervisor” is able to handle enough vehicles simultaneously to make the business case work.

      2. Max Power

        We didn’t demolish the trolleys, they went bankrupt during the post-war boom when the public voted with their wallets to buy cars instead. Nowhere in America is public transit breaking even, let alone profitable.

  2. Es s ce tera

    To me the holy grail is this:

    “To put it another way, the only way self-driving cars appeared able to live up to their promise would be if all cars were self-driving and ceded control to a central network, so the network would control the behavior of all vehicles, greatly limiting hazards and random events.”

    Which we’ve not achieved yet. So, to me, to argue self-driving cars have failed or hit a wall
    because all cars haven’t been replaced with self-driving autonomous is…odd.

    Clearly we’re progressing toward it from a place where it was previously unthinkable and impossible, to now it’s clearly possible but is it feasible…

    Also, why are we citing a few accidents? Compared to the driving record for human beings?

    1. Yves Smith Post author

      Because these are not even remotely self driving cars. They all have human minders. So you have a doubling up of resources, not a reduction.

      And this article politely skipped over the many times self-driving cars have hopelessly bricked traffic by stopping in the wrong place, to the degree ambulances and police cars can’t get through.

      1. booze

        I’m not sure if you’re splitting hairs, but my understanding is that for the most part, Waymo’s cars do not have a person in them minding them. Perhaps you mean remotely?

        I’ve noticed that there was no vehicle operator in the front seat. Are some of your vehicles fully autonomous?

        Yes. Our Waymo One vehicles are fully autonomous. That means that on most occasions, no one is in the driver’s seat when you ride with us.

        1. ChrisPacific

          There have been genuinely self-driving cars operating on a small scale as an experiment in places like San Francisco. Broadly speaking it’s been an abject failure (no shortage of articles on it if you Google). They don’t seem to kill people at least, but they drive through coned off areas, ignore emergency services or manual traffic management, shut down traffic by stopping due to improper threat identification or network glitches, etc.

          A mere few hundred have clearly been a net negative for the city transport network, and if they were deployed at higher scale they’d likely shut it down completely.

    2. i just don't like the gravy

      You will eat bugs and be happy as your Human Capital Transport Tube shuttles you to your job at the arms factory as it coordinates – and surveils – the movement of all upstanding citizens.

      Don’t forget to say “I love Uncle Sam and artificial protein!” after you punch the clock, otherwise your credit score is docked points for unpatriotic noncompliance.

    3. John

      Not only do you have to control all the cars, you also have to control the pedestrians, the dogs and cats, the birds, wind storms, falling snow, etc. In the real world there are simply too many events (or to quote the article – tail events) that can not be predicted

    4. José Freitas

      This seems like a recipe for disaster to me. Suppose activists or a terrorist group takes over the control nexus… Suppose hackers target it for ransom… Suppose the government decides to suppress a demonstration, or decides to shuttle all traffic away from some dubious actions… Etc…

      1. lyman alpha blob

        For an interesting take on that scenario, I recommend the Terra Ignota political scifi series, which has exactly what you mention as a major plot point. And in series, the cars can actually fly! First two books are great, the last two not so much.

      2. cfraenkel

        The government? In today’s economic ideology? It’d be some sort of PPP, with the funds to pay off the award bribes to politicians coming from the monopoly bribes extorted to get ‘preferential’ deliveries through before your competitors. Car pool lanes will become ‘enhanced service levels’. Good, better, best monthly transit passes. You think the telco duopolies are greedy bad? Wait until there is no possibility of competition, your life is at stake, and you have no choice but to use the service.

      3. NYMutza

        How about simply coning the vehicles as has been done by activists in San Francisco? There are many avenues for sabotage.

    5. Not Qualified to Comment

      the only way self-driving cars appeared able to live up to their promise would be if all cars were self-driving and ceded control to a central network, so the network would control the behavior of all vehicles, greatly limiting hazards and random events.

      Don’t see how this avoids the fundamental problem of dealing with the unusual and unexpected, which a human driver has perhaps a second or so to recognise, assess and respond to.

      I recall long ago driving a county road after a very high wind which had partially brought down a power pole on one side of it, resulting in wires being stretched horizontally across the tarmac about four feet above it. Against a background of wind-tossed foliage the wire itself was invisible from any distance while I doubt any kind of radar or whatever these things use would have picked it up, or made any sense of it if it did. I only avoided driving into it because I saw the pole nearly down, knew from previous use of the road that the wires it supported crossed the road and extrapolated from that, that there would likely be wires somewhere they shouldn’t be, perhaps still live with 11kVa even if not potentially capable of taking the roof off the car and decapitating me and my passenger.

    6. fjallstrom

      If you are going to have a central controlled system, why cars? Trolleys, buses, subways and trains are much more energy efficient, takes less space and are easier to schedule. Cars advantage is that they are individually controlled motor vehicles that carries up to a family unit and lots of stuff.

      And of course in a centrally controlled system you can’t have exceptions, like old cars, motorbikes, bikes or pedestrians, completing the cedeing of the roads to (centrally controlled) cars.

  3. upstater

    Note how the quoted excerpts repeatedly uses the word “accident”. An accident is an unforeseeable event in the legal sense. Because a human operator would recognize a hazardous situation and the software does not, the autonomous vehicle incidents are not “accidents”. The word has be misused for a century not to describe the carnage on roads; media even uses it to describe DUI deaths which are completely foreseeable.

    1. Socal Rhino

      Covered in the movie Hot Fuzz: Why do we call it a crash? Because accident implies no one was at fault.

      1. Nordberg

        I am glad to see a Hot Fuzz reference. I may need to queue that up to watch over the Holidays. Thanks for the reminder!

  4. Opusonthemove

    Long time reader, have been taking a sabbatical and travelling for the last year, driving from the USA to South America. Considering autonomous vehicles for the rest of the world as we have been driving is a fun topic of conversation. Observations: in comparison to Latin America, the driving environment in the USA is already incredibly simplified. The number of incomprehensible situations, intersections, and driver decisions we encountered every day is absolutely mind boggling. Using a cell phone while driving is simply not possible because you must be ready for insanity to break out at any moment. Unlike other technologies (cell phones & 4G data for example, which we have found to be far better and 20% of the cost of the US) that are easily adoptable in developing economies, I see no path for self driving vehicles anytime soon in much of the world.

    1. Expat2uruguay

      Absolutely true. In the US everyone owns a car so there are not very many buses, or motorcycles, or bicyclists, or pedestrians relative to Uruguay and other South American countries. So the talk of a network that controls all of the driving vehicles being able to avoid hazards of unknown Origin completely ignores the rest of the world that isn’t inside a car.

      In fact I find all the commenters here seem hopelessly wedded to a driver-only perspective and I wonder, where are the bicyclists, motorcyclists and the pedestrians in our comment section? We are the ones who will be maimed or die so that other people can watch tiktok videos during a short trip that they drove instead of walking in the first place. Where are the people who want walkable cities?

      All of this is so hopeless! Cars are not the future at all. We did not evolve to sit around all the time and it is killing us. Get out of your cars!!!!

      1. Expat2uruguay

        I forgot to add, that here in Uruguay I still see horse drawn carts that are used by the garbage recyclers! On the roads, of course, mostly in the neighborhoods, but also on more busy streets

      2. NYMutza

        The entire US lifestyle is absurd. Suburbia is absurd. Interstate highways are absurd. Single occupancy vehicles are completely absurd. And yet, here we are. A pretty hopeless situation.

      3. Uncle Doug

        Hear, hear!

        I am among those who have been fighting, and losing, the battle against American Vroomism for decades. In only a relatively few US locations have even modest efforts to tame carmageddon and improve the streets for non-motorized users been significantly successful.

        In the small city where we live in retirement, a “complete streets” project was killed after design and funding, by politicians who ran for office on a platform of killing it. We heard campaign statement s by people railing against the possibility that the new designs would force drivers to slow below the speed limit to accommodate turning traffic or crossing cyclists and pedestrians.

        The American infatuation with the private automobile is utterly bizarre.

  5. The Rev Kev

    They were never going to work. Silicon Valley simply decided that they were going to send those cars out onto the roads and they used their billions to ignore laws and to buy off local politicians as well as those in the State and Federal government. But the software was never up to the job. In fact, it was beta software and they were going to refine and upgrade that code by sending those cars out into the real world to see what would happen. This was the Microsoft approach where they shipped their buggy beta software and let all the users be their free fault reporters. Only difference is that if Microsoft Windows crashed, normally people would not die. But that is the virtual world and here we are talking about cars in a real world setting. So if people died from time to time because the software was not really going to work, well, it was a price that they felt worth paying as far as Silicon valley was concerned. Looks like it was reached critical mass though where even the techies have been forced to admit that this generation’s code is not fit for purpose.

    1. NYMutza

      The US presidential candidate-in-waiting Gavin Newsom knows which side his bread is buttered on. He recently vetoed legislation that would have curtailed driverless big rigs operating in the state of California. This was done clearly at the behest of Big Tech.

  6. Acacia

    one is still left wondering, how did such a not-ready-for-prime-time technology get so far?

    There is a widespread misconception that great progress is being made with this tech, very rapidly.

    But in fact, if you take the long view, the problems of computer vision and autonomous navigation in real world environments remain extremely difficult.

    In the 1980s, I worked in an academic lab alongside a number of researchers in vision and robotics at Stanford. At one point, some of the grad students screened a film that had been made about ten years earlier, to demonstrate the then state of the art. Watching this film, what struck me was how similar the tech was to their current work, showing just how little progress had been made in over a decade of research. But, the DARPA grant money had to flow.

    Since those years, I’ve always felt deeply skeptical about the claims for AI and autonomous vehicle tech. The current implosion of the hype thus comes as no surprise.

      1. Acacia

        Understood, and it has been rather surprising to me that this was ever allowed. Aside from the points raised in the article — e,g., that the complexity of the real world has been drastically underestimated —, though, I would still submit this has happened in part because many people believe that great progress is being made with the technology and it’s just a matter of shaking out a few bugs.

    1. Paul Art

      Agreed. Spent 4 years (2019-23) at a company that makes Radar for automobiles. I wrote signal processing algorithms that calculate range and range-rate of targets. Technology has not come to the point where object identification itself is quick and smooth. They are trying to use AI and ML (Machine Learning) now but I would say we are still too far off from actual autonomous vehicles. Detecting slow moving objects like pedestrians and also debris on the road are still challenges. Everything depends on the leaps we make in signal processing and so far the progress has not been spectacular.

  7. Martin

    Maybe it’s just that I’m from south america, but I can imagine instances where a driverless car “safety first” logic would be abused. This is why I believe the greatest challenge will not be technical, but a social one.

    How will these cars not be stolen and stripped for parts on a regular basis? When stealing a car, usually thieves would have to deal with a driver who can run them over if being carjacked, and with other people around who may intervene. With a self-driving taxi, they can prepare the whole operation, call the taxi and have it arrive, and then just put a tarp over it and do whatever they wish with it. I would guess it can report itself stolen and sound an alarm, but I’d be surprised if that would achieve much.

    Other than that, I can imagine self-driving taxis being a favourite tool for protesters, who can halt traffic just by standing in front of them, or damaging their sensors. I can also imagine people not giving a damn and crossing the street while self-driving cars slow down and stop so as not to hit the pedestrian.

    1. Henry Moon Pie

      That’s what the robot dog in the back seat and the Aston Martin DB5 machine guns under the headlights are for.

    2. Uncle Doug

      “I can also imagine people not giving a damn and crossing the street while self-driving cars slow down and stop so as not to hit the pedestrian.”

      I vow to do exactly that, whenever convenient, right here in the US, whenever the presence of autonomous autos on our streets presents the option.

  8. Owen Finn

    When a paradigm shift away from our car-centric cities and lifestyles is needed, the hype about driverless cars only serves to reaffirm the present paradigm. More people living in walkable, bikeable, human scale places with plenty of trolley and train options will revitalize human connections and help solve a plethora of social, economic, and health problems.

  9. TomDority

    Promoting Self-driving cars has been great for stock investors – after all – you have to monetize and maximize your take before the escape

  10. Mikel

    “In 2010, at the Shanghai Expo, General Motors had produced a video showing a driverless car taking a pregnant woman to hospital at breakneck speed and, as the commentary assured the viewers, safely….”

    Did they stick lights and sirens on the car?

    Speaking of sirens, that’s just another reminder how many of the senses are needed when driving. Our brains operate them simultaneously, not one event at a time.

  11. Brian Westva

    I’m glad to hear that driverless cars are most likely going to be unsuccessful. When I drive in large metro areas, I think that the driving test should be much more difficult and car owner much higher to limit the number of people driving badly on the highway. Driverless cars would only add to the problem of bad human drivers.

    Maybe we should go back to the original driverless cars……….horses!

    Unfortunately, I can envision this driverless technology being used by the military in their future wars. In that scenario hitting people and breaking stuff would be a feature not a bug. A remote controlled tank or fighting vehicle will allow the military to lower its standards. Don’t need to be fit if you are just using video game controllers. Russia already has a robotic mini tank.

    1. LifelongLib

      The last time I rode a horse I had the same thought — this is what a truly autonomous vehicle would be like. On level ground it more or less did as I directed but on a slope it ignored me and found its own way. Quite a different experience from a car.

  12. Michael.j

    Your long term studies of driverless cars have always fascinated me. Scanning the tubes recently I noticed that they have “taken over” Beijing.

    Has anyone else read anything about this? I wonder if the economic system there is effecting the quality of the software.

    1. CA

      Your long term studies of driverless cars have always fascinated me. Scanning the tubes recently I noticed that they have “taken over” Beijing.

      Has anyone else read anything about this?

      [ Driverless vehicles, delivery vans and taxis, are indeed seen all over Beijing. The Chinese media report this frequently. I have no idea why notice has not been taken abroad:

      November 11, 2023

      Driverless delivery vehicle to steal the thunder of “Double Eleven” shopping festival ]

      1. loneplateau

        The article indicates that in a city of 1.3 million people 12 driverless delivery vehicles are operating. Hardly a takeover.

      1. Michaelmas

        Interestingly, in the Reuters article’s picture of the driverless personnel transport car in Shanghai, those look like the two towers of the Fangjiashan Nuclear Power Plant, which are just outside town and were built in 2008-9.

        And the Chinese have just cleared their first flying taxi for business —

        As with when Silicon Valley cryptocurrency companies made claims for how their product would enable people to send people money between banks faster, etc. etc. — in other words, do all the things that standard electronic banking and contactless networks had already been doing for decades in places like the EU and UK — William Gibson’s dictum about the future being already here but unevenly distributed comes to mind.

        Sure, autonomous vehicles are hard. But a lot of it is that in the US these kinds of new technologies are principally scams to secure VC investment, whereupon they’re poorly implemented and rushed to market before they’re ready, so they then fail.

    2. CA

      Evidently, this fully-driverless taxi experience has been successful:

      March 27, 2023

      Driverless cars ready for hire in suburban Beijing
      By Gong Zhe

      Driverless taxis are now available for hire in suburban Beijing, marking a new chapter in intelligent transport, according to media reports on Monday.

      Beijing earlier in March granted licenses to Chinese tech giant Baidu and Toyota-backed autonomous vehicle startup to run fully driverless robotaxi services.

      This is the first time that a fully autonomous fleet of vehicles has been granted permission to operate in a world-class metropolis, the reports said.

      Previously, robotaxis were required to have a human driver in case of an emergency. However, the new permission will allow the two companies to operate robotaxis with just the passengers inside….

      1. Stubbins

        I would expect China to allow a pilot-less flying taxi with even more lax permissions than the US, but after the inevitable accident where it takes out a crowd of people, to also sentence the grifting CEO to death.

  13. Cynical Engineer

    In the software world, there’s an old joke:

    The first 95% of the work takes 95% of the time. The last 5% of the work takes the other 95% of the time.

    Autonomous self-driving cars using trainable neural-networks solved the first 95% of the problem fairly rapidly, a decade ago. They’ve been struggling to master that last 5% ever since, with shockingly little progress to show for 15 years of effort and $Billions spent.

    The core problem is the difference between rote memorization and true understanding. It’s utterly impossible to rote-train these systems on every possible situation they can encounter. The underlying system is not “intelligent” and capable of understanding the problem in a human fashion. Until that future day when a breakthrough gives computers true human-style “intelligence”, the computers will not be able to handle tasks of this complexity.

    The people involved have been confidently predicting they would “solve” this next year….for the last 15 years. At some point we need to acknowledge that it isn’t happening and move on.

  14. Cynical Engineer

    The other aspect of the self-driving cars that has been ignored is economic:

    The computer equipment need to make this happen is large and power-hungry. It takes up most of the available cargo capacity in the current models.

    When combined with the required sensors, your average “Self-driving car” costs between $150,000 and $250,000….FOR EACH CAR. And that doesn’t include garage/maintenance facilities, or the people to staff them. (Tesla’s “Autopilot” system is much, much cheaper. It’s also spectacularly bad.)

    When you do a business plan based on current self-driving cars, and compare it to a traditional taxi service, the numbers don’t work. The traditional taxi is massively cheaper to operate.

  15. Glen

    Tesla might have a bit of AI implemented in it’s cars:

    NHTSA Finds Teslas Deactivated Autopilot Seconds Before Crashes

    A NHTSA report on its investigation into crashes in which Tesla vehicles equipped with the automaker’s Autopilot driver assistance feature hit stationary emergency vehicles has unearthed a troubling detail: In 16 of those crashes, “on average,” Autopilot was running but “aborted vehicle control less than one second prior to the first impact.”

    AI right before it hits the cop car, “Oops, you got it buddy!”

  16. Roger

    As a person who spent his whole career either coding or managing software development teams self-driving cars never made sense. Even in banking systems the vast majority of the code is there to deal with exceptions to the standard, and that is with quite a controllable problem space. The autonomous vehicle will exist in a vast ill-defined problem space which simply cannot be coded for or “learnt” (all the AI hype is really about enhanced learning algorithms).

    The greater negative outcome from a mistake made by a system, for example human injury or death, the lower probability of failure there has to be. Its is well known that the last 10-20% of the problem space will take up the most code and that relationship become greater the closer you get to 100% (and 100% can never be attained). Then on top there are all the legal issues of whose fault a crash will be (the driver., the software company, the vehicle manufacturer, the sensor manufacturers?).

    I for one will be very happy for “self-driving” to be given the proper label of “assisted human driving” and Musk to have to stop mis-selling his FSD product (there should be a massive consumer lawsuit there once Musk’s halo has dimmed). Also all of the “sharing” apps which are really just ways for rentier-capitalists to destroy well regulated industries and exploit drivers while hubristically breaking the law (and burning huge amounts of investor cash). The world we be a better place (as long as municipalities remember that taxi licenses are their property not that of license-hoarding exploiters).

  17. Joe Well

    The real value of self-driving car hype was in denting the momentum for major investment and innovation in public transportation and other shared/communal transportation. The US totally ceded that to China and other countries.

  18. Simple John

    Just because you can name something doesn’t mean it has any claim on existence. Elon named a pet hobby and the world suspended judgement.
    I came to Tesla late but I thoroughly enjoy driving my torque in a box.
    AI in general makes people dissociate. Think about it.
    Think of what people know that they don’t have words to say.
    And think of what people know when recreational drugs kick in.
    Believing that AI will replace brains is like believing Legos and Lincoln Logs will be the building blocks for the world’s housing and industry.

Comments are closed.