Doubts About Self-Driving Cars

A new poll finds that nearly two out of three auto owners think self-driving cars are a dangerous idea. Slate writer Lee Gomes argues that self-driving cars may never happen. Both are wrong.

The pollsters don’t argue that self-driving cars actually are dangerous; only that “automakers will have to work to win over car shoppers who think some of the technology makes vehicles more dangerous.” But they really won’t; they just have to make the technology available to early adopters, and as those pioneers prove it to work, more people will want it.

Gomes’ argument is that Google’s self-driving car critically depends on accurate maps, and such maps are expensive and time-consuming to make. Moreover, Gomes adds, keeping the maps up to date with daily changes in routes, traffic signals, speed limits, and other factors will be nearly impossible.

It provides erection viagra soft 50mg that stays for longer time duration. Therefore whenever your impotency affected body wants to avail price levitra back its stamina and strength you can impute this medicament in order to get back your sexual strength. Several tests are needed levitra without prescription to find out this sexual disorder face enough mental traumas and can realize the results. Firstly, in tab sildenafil order to remove the nofollow tag then there are different methods for different platforms. Writing in Forbes, economist Adam Ozimek responds to Gomes by noting that the states themselves are likely to make maps and will keep them up to date. In fact, many government agencies are already using LIDAR (the same technology used by Google) to make maps. The data these agencies generate will be public information. It seems likely that someone will start a business of simply gathering all public LIDAR data and making it available to software companies to include in their autonomous vehicle maps.

In addition, Gomes fails to take Moore’s Law–the notion that the amount of computing power and memory that can be purchased for a fixed number of dollars is growing exponentially–into account. Thanks to this computing power, map making is getting less expensive and the cost of including such maps on your computer is declining as well.

Beyond this, Gomes makes the mistake of thinking that self-driving cars won’t be able to function without perfect maps. While maps are the starting point for autonomous vehicles, they will not be the only tool used by such cars. Instead, the cars will rely heavily on sensors that can detect and deal with exceptions to the maps.

Only time will tell for certain whether fully self-drivable cars will be on the market by 2020. But anyone who is counting on them to fail is likely to be disappointed.

Tagged , . Bookmark the permalink.

About The Antiplanner

The Antiplanner is a forester and economist with more than fifty years of experience critiquing government land-use and transportation plans.

20 Responses to Doubts About Self-Driving Cars

  1. FrancisKing says:

    “But anyone who is counting on them to fail is likely to be disappointed.”

    I think that the opposite is more likely.

    From year dot, car drivers have had this vision of car-driving Utopia thrust upon them. An unattainable vision – but if only ‘_____’ happened, then the vision can still be delivered. Pollution from cars can be fixed by hydrogen powered fuel cells – but please don’t ask who’s going to pay for the entirely new fuel delivery system. If only we could get cyclists, buses, pedestrians out of the way, the roads would be clear for cars – but please don’t ask what the cyclists, bus passengers and pedestrians are going to do instead. Safety and road capacity can be fixed by automating the cars – but please don’t ask how this is going to happen, when Google have already stated that their cars will come programmed to drive exactly as car drivers do – including speeding and delayed braking.

    I can’t see many people wanting to pay for this, unless perhaps they drive long distance, where the automation can function as a ‘co-driver’. The neat thing about cars is that the driver comes for free – a basic point undermined by the automation process.

    There are some really neat options out there. For example, a large proportion of the cost of running buses is the driver, but during peak periods, when we really need buses as a way of reducing car flows, half of the time (out of town in the AM. into town in the PM) the buses are empty or nearly so. Automation of buses means that a member of the public can drive the bus one way, and the automation takes it back. That’s a lot more interesting.

  2. Fred_Z says:

    “the states themselves are likely to make maps and will keep them up to date”

    Sure they will, perfectly too, with never an error, just like they do now with all of their programs. Ha, ha, ha. For a man with little faith in government, the Antiplanner sure has a lot of faith in it for all things supportive of self driving cars.

    The Antiplanner also misunderstands Moore’s law. It’s not a law, it’s a theory, and even if it continues for a while it cannot continue indefinitely. The Persian Prince, the chessboard, doubling the grains of wheat and all.

    In any event, human programming skills and quality are not progressing at the same rate, and it is the software that will be incredibly difficult. How many Microsoft bug patches have you been forced to download this month?

    I have 27 for the first 17 days of November alone, starting with Definition Update for Microsoft Security Essentials – KB2310138 (Definition 1.187.1137.0) 2014-11-02 and ending with Definition Update for Microsoft Security Essentials – KB2310138 (Definition 1.187.2344.0) ?2014-?11-?16

    It is a poorly kept secret in programming circles that all code is incredibly buggy, not getting particularly better, and often getting worse. A poor programmer often makes code worse in attempting to fix bugs.

  3. Frank says:

    I see the technophobes have recycled their same old unsupported assertions.

    Fred_Z engages in faulty reasoning, even though evidence has been provided that autonomous systems have redundant backups. His pointing out imperfections in code is quite ludicrous. Somehow airplanes, which are highly computerized, manage to fly. And of course, why doesn’t he point out other imperfections in cars? Certainly tires fail, causing crashes. Radiators. Belts. All kinds of technology fails. Yet Fred is fixated on programming.

    “How many Microsoft bug patches have you been forced to download this month?”

    None. Because I no longer use that shitty OS. Not sure why anyone does. I’m an somehow able to type this comment on Chrome OS, and I haven’t downloaded ANY security patches, ever! Thank god Microsoft isn’t developing an autonomous vehicle.

  4. bennett says:

    I share Mr. O’Toole’s optimism about self driving cars. I think that the fear of letting go control of your vehicle will pass with time. I also share in the trepidation of such systems, but I mostly recognize that it’s because it’s a big change and it’s “new.” Big new changes are kind of scary, but we’ll get over it.

    I also sympathize with Fred_Z’s comments about computing and programming. I think to myself “what could be scarier than a computer operating a vehicle?” Then the answer hit me like a ton of bricks… “A Texan operating a vehicle!”

  5. Frank says:

    Then the answer hit me like a ton of bricks… “A Texan operating a vehicle!”

    I sincerely lol’d at this. You can also replace “Texan” with “stoner” or “drunk” or “hipster” or “teenager” “soccer mom” or …

  6. bennett says:

    Frank,

    Every one of those people live in Austin and they’re all Texans. The roads in TX, particularly in Austin, Dallas and Houston are a hellish nightmare in which people are not actively driving despite the fact they are behind a wheel and moving a tremendously high speeds, unless of course the are in the left lane, in which case they are talking on their cell phone and going 10mph under the speed limit. My palms are sweating just typing this.

  7. prk166 says:

    When day in, day out most of deal with PCs that randomly die, have weird errors, constantly require reboots, get hacked, perform worse than they did a few years ago when we bought it new, it makes it hard for us to imagine computers becoming good enough and reliable enough to do something as complex as driving a car. I get it. I have my moments of doubt, too.

    The map comments remind me of something similar. A decade ago had you asked people about Google providing real time traffic information on 3/4th or more in all sorts of cities – sometimes even small towns – across they USA they would’ve scoffed. It would’ve been far, far, far, far too expensive for them to go out and install the millions of traffic cameras that it would’ve required. Imagine the expenses involved in wiring the whole thing up on top of staffing the operations and there’s no way even a giant company with fat profit margins like Google would ever do it. They may have even laughed in your face had you insisted it would happen.

    Guess what? Google does it today. On top of it others do it too!

    The key is they didn’t do it by creating a system using the technology of that day. They didn’t get stuck in the mud, only able to imagine doing it the way state DOTs and regional and city governments were doing it. They realized that once enough phones were in play, people would give them permission to use the data from their phone to determine traffic levels.

    Maybe we’ll never have driverless cars. Maybe it’ll be too complex. But as the above and countless other innovations throughout history have shown us, we shouldn’t be quick to scoff.

  8. The difference between PCs that randomly die and computers in automobiles is that the former are general purpose hardware to which people add all kinds of software that does things the hardware manufacturers couldn’t and didn’t predict, while autos have processors that are specifically aimed at doing one function with one set of software. If you have a car less than 10 or 15 years old, it probably has a dozen more more processors in it; some high-end cars have more than 100. These cars don’t randomly die or require constant reboots.

  9. Here are some quotes from the story I cited in my previous comment:

    “Audi’s implementation of Google Earth and the Tesla Model S‘s 17-inch touch screen and instrument cluster are powered by Nvidia Tegra processors. The next generation of these chips will perform at 384 GFLOPS, or billions of floating point operations, per second. With four processors per car, a two-car garage would have as much computing power as the $120 million Blue Mountain supercomputer installed at the Los Alamos National Laboratory in 1998.”

    “Collision-avoidance image processing tops the list of data-intensive in-vehicle applications, requiring 100 billion operations per second.”

  10. Fred_Z says:

    Frank:

    Planes are indeed flown by computer, but they all have pilots. Google “autopilot malfunctions” and you’ll get over 67,000 hits.

    As for chrome OS, no updates? You gotta be kidding. see https://chromium.googlesource.com/chromium/src/+log/40.0.2214.6..40.0.2214.10?pretty=fuller&n=10000.

    The third commit there is for bug 433912. Four hundred and thirty three thousand nine hundred and twelve. Bugs. So far. Google is just less forthright about telling you they have had to fix over 400 thousand bugs. Google “Linux OS bug list” and you’ll get 67M links.

    The Anti-planner says “autos have processors that are specifically aimed at doing one function with one set of software. ” Ummm, so is Windows. So is Chrome OS.

    How many of you guys expressing faith in software actually write it? I do. I have been since I was a ‘maintenance programmer’ for Amoco using PL/1 in the seventies. maintenance programmers are the guys who get called in at 2AM on payday when the system is printing cheques for $0.00 or eighty million dollars. I have seen the best programs and programmers in the world do really dumb stuff. I have seen them miss terribly complex things despite their brilliance.

    I am anything but a technophobe. Love the stuff. I still program a bit. Wrote this little thingy this morning:

    function js_GetTable(pSourceOfCall)
    {
    var vParams,vReselectParams,vReturn,vSource,vTable;
    vReselectParams=js_GetReselectParams();
    vReselectParams.Request=’GetTable’;
    switch (pSourceOfCall.type)
    {
    case ‘click’:
    vSource=’from click.’;
    if (pSourceOfCall.target.parentNode.tagName===’DIV’)
    vTable=js_GetEl(gvTable1);//click is within a div, not table, so comes from a toolbar
    else vTable=pSourceOfCall.target.parentNode.parentNode.parentNode.parentNode;//click is within a table
    break;
    case ‘change’:
    vSource=’from automatic reselect.’;
    vTable=js_GetEl(gvTable1);
    break;
    default:
    vSource=’from reselect on login validation.’;
    vTable=js_GetEl(gvTable1);
    }
    if ($.inArray(‘Combi’,vTable.classList)!==-1)
    vReselectParams.Combi=’Yes’;//in a combi table
    else vReselectParams.Combi=’No’;//not in a combi table
    vReselectParams.TableName=vTable.id;
    js_AddMessage(‘ReSelect ‘+vSource);
    js_AjaxPost(false,vReselectParams,js_CallbackOnTableReplace);
    }//end js_GetTable

    It’s a javascript Ajax based HTML generator calling into PHP to access one of my personal MySQL databases. This silly little bit of code calls functions which call functions which call functions and so on nearly infinitum. Its proper working requires that probably five million lines of code work perfectly. They usually do. Until I try to run the same code in IE as opposed to firefox.

    Sorry for the long post. Good code, and its difficulty, are a bit of a passion.

  11. Frank says:

    The Antiplanner made the same point about personal computers that I was about to make. Also realize that PCs are hardware, with most of them running Microsoft’s bloated and crappy OS. And as the AP points out, they are multipurpose.

    Planes are indeed flown by computer, but they all have pilots. Google “autopilot malfunctions” and you’ll get over 67,000 hits.

    So? Number of hits are irrelevant. Plus you’re off topic. I never claimed that autopilot doesn’t malfunction; I did, however, claim that there are redundancies. Perhaps you’d like to come back with some better “evidence”. Also, if you limit your search range to results since 2009, you only get 308 hits.

    As for chrome OS, no updates?

    That’s not what I claimed, and I’m uncertain as to why you feel the need to strawman me.

    Here’s what I wrote: I haven’t downloaded ANY security patches, ever! And I haven’t. There aren’t “security patches” for Chrome OS, although there are updates to the OS, and they are automated, so I don’t have to manage them nor am I “forced to download” them.

    The Anti-planner says “autos have processors that are specifically aimed at doing one function with one set of software. ” Ummm, so is Windows. So is Chrome OS.

    Windows is aimed at performing many different functions (video rendering and playback, photo editing, web browsing, gaming, etc. ) with a variety of software, a great deal of which is third-party installed by users.

    How many of you guys expressing faith in software actually write it?”

    This is an appeal to personal authority (although I’ve been coding since the early 1980s when I used BASIC and currently code HTML, JS, and PHP—and have managed to get mostly out of the education biz and into IT, but still teach a coding class). I know all about bugs. But bugs haven’t stopped progress.

    Until I try to run the same code in IE as opposed to firefox.

    Why would you want to do that?

  12. Frank says:

    I’m not a blind believer in technology, but I am cautiously optimistic. The naysayers have always perplexed me; their crystal balls must be more polished than mine.

    For fun, here is a list of things expert naysayers have naysaid:

    “Everyone acquainted with the subject will recognize it as a conspicuous failure.”
    Henry Morton, president of the Stevens Institute of Technology, on Edison’s light bulb, 1880.

    “Fooling around with alternating current is just a waste of time. Nobody will use it, ever.”
    Thomas Edison, 1889.

    “We have reached the limits of what is possible with computers.”
    John Von Neumann, 1949

    “Transmission of documents via telephone wires is possible in principle, but the apparatus required is so expensive that it will never become a practical proposition.”
    Dennis Gabor, Hungarian-British physicist, 1962.

    “The ordinary ‘horseless carriage’ is at present a luxury for the wealthy; and although its price will probably fall in the future, it will never, of course, come into as common use as the bicycle.”
    Literary Digest, 1899.

    More here.

  13. metrosucks says:

    The real reason planners are afraid of self-driving cars, is because they are scared it will make their pet rail transit and mixed used boondoggles utterly, completely obsolete. Of course, they can’t say that, so they make up nonsense like demanding the software be 110% perfect before it is allowed in the wild. Why don’t we apply the same standards to their transit boondoggles and watch them get all quiet really quickly?

  14. bennett says:

    I agree with Frank… but there is a flip-side to that coin, one taken up by Mr. O’Toole on other posts re: automated cars. Yes, technological progress succeeds in spite of naysayers, but technology often fails to meet the goals of those who espouse it’s virtues. I like that Mr. O’Toole advocates for automated vehicles but I am doubtful that it will solve as many problems as he predicts.

    I always go back to computing and mobile/wireless technology. It was touted as something that would make us more productive thus giving us more time for leisure. The result have been an increase in productivity yet a decrease in leisure time. By making things “easier” it has made things vastly more complicated. Technological determinism always creates more problems than it solves.

    I also agree with part of mycommentssuck post. A fully integrated diverless car system would certainly undermine many pro-transit arguments. However, the implementation of such systems is not likely to result in any reductions in bureaucracies so planners have no need to worry.

  15. Frank says:

    Good points bennett. Yes, predictions, both positive and negative, just predictions.

    I sure agree with you about wireless tech increasing productivity but decreasing leisure time. Certainly the case in IT world, as well as business, education; you name it. It has also made things more complicated. In the last month, I’ve said many times that if time travel were possible, I’d go back to the 1980s before cell phones because America has become Zombieland, everyone glued to their devices and dining and walking (or driving) without paying attention to others around them.

    However, I’m not sure about your statement that “Technological determinism always creates more problems than it solves” especially since is a theory; are you stating that the implementation of new technology itself creates more problems than it solves? Either way, this is quite a generalization.

  16. bennett says:

    “…are you stating that the implementation of new technology itself creates more problems than it solves?”

    Not at all. In fact the opposite is usually true. What I saying is that philosophies about a given technology or technology in general that presume that a society’s technology drives the development of its social structure and cultural values, or put another way, a philosophy that seeks to show that a given technical development will be the key mover of social change, creates more problems than it solves.

    For example, if we are to assume that driverless cars will effectively solve congestion, reckless driving, auto fatalities, limits in productivity due to actively driving or using other transportation modes (which has been stated and alluded to on other Antiplanner posts) it absolves us of any responsibility to collectively solve these problems today. It is essentially my beef with Antiplannerism. We’ll just wait for technology, presumably created through free market enterprise, to work it’s magic. We don’t need to do anything (i.e plan), just go and by the new technology when it becomes available (because it’s obviously in our self interest).

  17. prk166 says:

    “The Anti-planner says “autos have processors that are specifically aimed at doing one function with one set of software. ” Ummm, so is Windows. So is Chrome OS.” ~Fred_z

    An operating system by it’s very nature does not do one function.

  18. metrosucks says:

    Most people think of Windows or Linux or whatever when they think computer. The systems in cars are more akin to PLC’s: rugged, reliable, efficient & heavily tested coding that only needs to perform its designed functions, and nothing more. Unlike a Windows PC, it doesn’t need to be compatible with 10 billion additional software programs or a million add-on hardware devices. There is almost zero similarity between a personal computer and the embedded systems that do the heavy lifting in critical areas.

  19. CapitalistRoader says:

    A new poll finds that nearly two out of three auto owners think self-driving cars are a dangerous idea.

    An old poll (from 1904) found that two out of three horse and buggy owners think that automobiles are a dangerous idea. Just guessing on this one.

    For example, if we are to assume that driverless cars will effectively solve congestion, reckless driving, auto fatalities, limits in productivity due to actively driving or using other transportation modes (which has been stated and alluded to on other Antiplanner posts) it absolves us of any responsibility to collectively solve these problems today.

    Sounds good to me. I envision urban US areas with millions of hyrbriid diesel (or natgas or hydrogen) self-driving cars, subscribed to and serving those with the highest subscription fees first, I and other cheapskates second, and parking themselves with when there’s no demand. And it will ultimately be individuals, not a collective, solving all the stated problems. It will be freakin’ wonderful, better even than the advent of the personal computer. And even more beneficial to the collective betterment of society than PCs.

  20. the highwayman says:

    The sophisticated technology that you want also makes humans obsolete. So Frank what are you going to do for money when you are replaced by a robot? :$

Leave a Reply