back to article Is Tesla telling us the truth over autopilot spat?

An increasingly bitter fight between Tesla and its former autopilot partner Mobileye is raising questions over the electric car company's honesty. Elon Musk's corporation is notoriously prickly. When it does come under criticism – whether on safety issues, the practicality of long-distance drives, or its autopilot feature – …

  1. captain_solo

    The whole reason the OEM auto companies have been so much slower than Tesla on many of these technologies is that they have a century of experience integrating technology into vehicles and testing and developing them over generations of model designs to ensure safe and reliable operation as well as recognizing that if they put tech into a vehicle, it could be on the road for 20 years. A time scale Silicon Valley never has to think about. Certainly there are cost considerations, obviously Tesla could not take over the whole auto market at their price point, the OEMs have to consider their entire product line including vehicles whose profit margins are razor thin.

    Obviously the auto companies aren't always great at getting this right, but they certainly know how hard the regulatory and public opinion hammer will land on their heads if they play fast and loose with safety in a life-critical application like operating several thousand metal and plastic parts flying in relatively tight formation down the highway. They are much more risk averse than Tesla and the reason's why are starting to become apparent.

    Driver assist systems when combined with an experienced driver make the car vastly safer, but trying to take the human wetware out of the equation at this point is premature.

    1. bazza Silver badge

      Well put. But never mind Tesla and self driving cars in general, the OEMs have cocked it up already with simple things that have nothing to do with self driving.

      The OEMs are rapidly discovering the costs of putting in even quite "simple" things like cellular modems and Internet connectivity into their cars. In the good ole' days they'd design a car, shove a bit of firmware here and there, sell it, forget about it (for the firmware will probably work trouble free for the lifetime of the car), move on to the next model design.

      But now there's an Internet connection, suddenly they've become a software company with a permanently stood up dev team dealing with bug fixes, vulnerabilities, patches, updates, maintenance and compliance testing lasting years and years and years with a the burden of public expectation being far higher than that experienced by, e.g. Microsoft or Apple. The fact that they install their software in something that has an engine and four wheels is almost a passing consideration in comparison. And these devs aren't even working on new models, they're looking after the ones that already out there. Expensive!

      And, as you point out, if they're ever caught out or get it wrong at any time over the next 20, 30, or 40 years then there may be hell to pay at worst, or severely pissed off customers at best. Fiat-Chrylser copped a half billion dollar fine for not addressing a trivial vulnerability in a software system that had limited real world utility for the customer/driver, and that didn't even kill anyone. Toyata, with their dodgy accelerator pedals a few years back, lost an absolute ton of market share and customer goodwill; imagine the reaction if a software fault / vulnerability was exploited to produce a similar effect.

      At the Goodwood Festival of Speed once I saw a team trying to get a 10 year old F1 car running. They really struggled, and it was all software and control laptop antiquity. Ten years old, took them a few hours to work out how to get the damned thing running. What hope for Joe Public when their 10 year old Tesla/BMW/Audi doesn't work one morning?

      1. Anonymous Coward
        Anonymous Coward

        Re Toyota uncommanded acceleration

        "Toyata, with their dodgy accelerator pedals a few years back, lost an absolute ton of market share and customer goodwill;"

        Toyota's issues go further than dodgy accelerator pedals and lost goodwill. Maybe you knew this, I'm not sure.

        For example, there was a $1billion penalty payable by Toyota US in order to shut down a criminal investigation:

        https://www.washingtonpost.com/business/economy/toyota-reaches-12-billion-settlement-to-end-criminal-probe/2014/03/19/5738a3c4-af69-11e3-9627-c65021d6d572_story.html

        In the court case which ultimately led to the $billion settlement (Bookout v Toyota Motor), evidence emerged to suggest that Toyota's design practices (hardware and software) for safety critical control systems left a great deal to be desired:

        Short version: http://www.eetimes.com/document.asp?doc_id=1319903

        More detailed version from Prof Phil Koopman at CMU who was an expert witness at the trial:

        https://users.ece.cmu.edu/~koopman/pubs/koopman14_toyota_ua_slides.pdf

        Share and enjoy.

        1. Cynic_999

          Re: Re Toyota uncommanded acceleration

          I just read that PDF. Good Lord! Apparently Toyota's code had a watchdog timer that was reset *from a timer interrupt routine.* That completely defeats the object of having a watchdog - the main system could hang completely, and the timer interrupt would faithfully keep ticking away, ensuring that the hardware watchdog never noticed. It would be a pretty unforgivable schoolboy error on the firmware running your TV set, but on a safety-critical system it amounts to designed-in criminal negligence.

    2. Anonymous Coward
      Anonymous Coward

      captain_solo, your comment is sort of bullshit. It reads as if it is written directly by General Motors public relations department. It would be complete bullshit if you didn't try to appeal to reason, so good job there.

      Go on, shadow Tesla in that uncertain light of certain doom. But last I checked, when _1_ person dies in a car due to design fault and the manufacturer takes _IMMEDIATE_ attention to that death, that was a good thing. Sorry, I just never got used to the idea of "Well, if enough people get hurt, we'll take a look ;-P".

      I really don't care about the automotive industry, but nearly ever American my age has heard of Preston Tucker. Big automotive is doing to Tesla what they did to Tucker over 60 years ago. Fuel the FUD, get government (now governments...plural) on your side, then steal the ideas and claim it as your own. Big automotive admitted to the scandal with Tucker AFTER they destroyed him, then later convicted of the plot. Like it or not, Tesla is the new Tucker.

      So, will I buy a Tesla, no. Will I _ALMOST_ blindly support Tesla in every other way, you're god damn right I will. Especially when I'm _FORCED_ to have my taxes support this apparent "Too big to fail" industry made of scum and vinyl seats (vinyl might be too good of a description).

      Drive safe.

      1. captain_solo

        Your taxes are supporting Telsa too in a big way. Don't be too quick to make them so different from the rest of the corporate piglets hanging off the Gubmint's teats. GM is such an easy target, I get it, and I agree largely, but come on. Tesla is just as bad or worse on the taxpayer funded boondoggle scale.

      2. Anonymous Coward
        Anonymous Coward

        Oh Dear

        "So, will I buy a Tesla, no. Will I _ALMOST_ blindly support Tesla"

        Sigh. I suppose you think you are the voice of reason?

        1. Anonymous Coward
          Anonymous Coward

          "Immediate attention"

          Is that why news of the crash only became public two months after the crash? If Tesla had their way they would have hushed it up forever. Even now they haven't learned their lesson, and still allow people to "drive" hands-free when they admit the software is not able to handle all situations.

          More people will die before they come down off their high horse, the question is whether the company will be able to survive it.

    3. Anonymous Coward
      Anonymous Coward

      > The whole reason the OEM auto companies have been so much slower than Tesla on many of these technologies is [...]

      Could you please share the source of that assertion? Or is it your personal opinion? In that case, I would appreciate a little (anonymous if you wish) background on your auto industry credentials. Cheers.

      1. captain_solo

        I have had some of these companies as my customers and while I don't have specific knowledge of this particular design path, i'm asserting it based on my experience with their comfort level with risk and liability.

        1. Anonymous Coward
          Anonymous Coward

          > I have had some of these companies as my customers

          So it's your own speculation then. While that's entirely respectable, I suggest it may be helpful to make that clear in your posts, for the benefit of that part of the audience who are primarily interested on evidence-based insights.

      2. big_D Silver badge

        I can remember watching Tomorrow's World in the 70s and 80s and seeing self-driving experiments from the likes of Volvo, Mercedes and VW Group being shown on a regular basis. The major manufacturers have been experimenting for over 40 years on building in automation to vehicles.

        Some of it back then was pretty basic, I think the VW system was mainly aimed at slaving the following cars to the leader on the autobahn.

        But if you look at what has trickled down to today's cars, you can see that some of that technology is slowly maturing enough to be used safely on the road, as assistive technologies. The ADAC (German motoring organisation similar to the AA in the UK or AAA in the USA) did a test last month of the assisted braking systems in BMW, Mercedes, Volvo, Kia, Subaru, VW/Audi and a couple of others. What was interesting was, that none of these systems were 100% effective in avoiding an accident, in fact the ones you would expect to be good (Volvo, BMW and Mercedes) faired very badly in the test. Unfortunately, they didn't test the Tesla S, so I don't know how it would have faired.

        But only two cars avoided colliding with a slower car in front of them - most braked hard, but still rolled into the "car" in from (a bag full of air, shaped like the back of a VW Sharan). When it came to pedestrians, again only 2 of them completely avoided hitting the pedestrian, and only the Subaru managed to stop completely at night, many of the others didn't even try and slow down! And none of them could cope with a cyclist - some slowed down, but all of them hit the cyclist to varying degrees (from around 20% speed reduction to 85% speed reduction from 50km/h). Unfortunately, i couldn't find an online link to the test - it was in the member's magazine.

        1. bazza Silver badge

          @Big D

          "The ADAC (German motoring organisation similar to the AA in the UK or AAA in the USA) did a test last month of the assisted braking systems in BMW, Mercedes, Volvo, Kia, Subaru, VW/Audi and a couple of others."

          Thatcham here in the UK put out a press release a few years back about their plans for how they'd test these sorts of systems. It was quite interesting.

          I took a glance at it, one thing that stood out was that all their tests focused on determining that a car does actually stop when it should do in a variety of scenarios.

          They were quite proud of their tests, but there was a glaring omission. How about proving that the car doesn't stop when it shouldn't?!

          That's just as important. For example, you really, really, don't want your car getting spooked by flying leaves, rubbish, etc. when you're heading down a busy motorway. Yet none of Thatcham's tests were going to explore that important aspect of the behaviour of these systems. Nor was it clear that there'd be anyway for a driver to prove that this is what had happened, so when it does it'd be the driver that takes the blame. Great. Basically, it was a set up to ensure that the insurance industry's interests were looked after, but not the driving public's. Still, that's Thatcham's job I suppose.

          I concluded that a dash cam would be an essential component of a driver's equipment to help act as independent corroboration of claims that such a car had taken inappropriate action all by itself.

          1. JetSetJim

            @bazza

            > How about proving that the car doesn't stop when it shouldn't?!

            Give them a complete list of these scenarios and the test conditions then. I suspect that's why that wasn't in their list, as I'm sure it would be all too easy to miss a few dozen and then Thatcham would have egg on face. They should at least mention it, though, and I bet their fine print in whatever report they produced would do so.

          2. big_D Silver badge

            At the weekend, I noticed that my Qashqai was beeping at me at junctions, as though something was approaching the front end at low speed (front parking sensor). There was nothing, I assume it was residual water running over the sensors when braking. Luckily it didn't try and slam on the brakes, but it was a little annoying.

          3. Anonymous Coward
            Anonymous Coward

            Re: How about proving that the car doesn't stop when it shouldn't?!

            Been there, seen that.

            I ended up owning a small city car which came with direct sequential auto gearbox and city brake assist (which is reliant on some kind of forward facing sensor).

            Every now and again while driving along it would go into neutral, for no visible reason. Sometimes a warning light might come on, but the dealer never found anything in the logs they could see.

            Over a period of months of low speed commuting (five miles a day, half an hour a day) I had failures every few days becoming increasingly (but never 100%) repeatable once I worked out what was most likely to make it happen.

            Eventually one of the control units was "updated" by the dealer.

            And the problem went away.

            1. Anonymous Coward
              Anonymous Coward

              Re: How about proving that the car doesn't stop when it shouldn't?!

              > Every now and again while driving along it would go into neutral, for no visible reason

              Some VW cars have a "coasting" mode which will do what the name says: shift into neutral when the engine is not actively pulling the vehicle and no engine braking is required. I suppose that wasn't the problem, but thought I would mention it anyway.

              1. Anonymous Coward
                Anonymous Coward

                Re: How about proving that the car doesn't stop when it shouldn't?!

                "Some VW cars have a "coasting" mode which will do what the name says: shift into neutral when the engine is not actively pulling the vehicle and no engine braking is required. I suppose that wasn't the problem, but thought I would mention it anyway."

                Oh. I didn't know that. Thank you. And I'm not sure the dealer knew it either (it was a VAG dealer I'd known for years and would trust with old-skool stuff. Not so sure about computer stuff).

                "Coasting, with no engine braking required" might even be consistent with the circumstances which seemed most likely to provoke it. Interesting. Is there somewhere I can read more?

                fwiw It was Skoda's equivalent of the VW Up! and was otherwise quite a nice car for round-town commuting, though the first time this symptom showed was on a slight bend at the bottom of a dip in a country road, and many other occasions involved something which I suspected the car might have incorrectly detected as an obstacle (e.g. parked vehicle on exit from a roundabout).

                No excuse for going into neutral and staying there though.

                1. Anonymous Coward
                  Anonymous Coward

                  Re: How about proving that the car doesn't stop when it shouldn't?!

                  > Oh. I didn't know that. Thank you. And I'm not sure the dealer knew it either (it was a VAG dealer I'd known for years and would trust with old-skool stuff. Not so sure about computer stuff).

                  To be fair, that's a user configuration setting (meaning, the driver changes it from the console with the car stopped), so there is a good chance that the dealer did not think about it, expecting to find an actual fault instead.

                  As I recall there was a section in the manual. See if you can Google up a copy for your model, or borrow one from someone else. The gist of it was that it would go into neutral if you let go of the gas and did not touch the brake, if cruise control wasn't enabled (I think), and if the gearbox was not in "S" mode.

      3. c1ue

        I used to support Motorola - including their automotive division. It would take 4 years for a new system to make its way from fabricated and lab tested system to installed in any actual consumer bought vehicle.

        A car is not a great environment for electronics. There is tremendous vibration, there's a lot of dust and other foreign objects, huge temperature swings, and significant ESD. With an electrical car, you likely also get significant magnetic field perturbations due to the electric motor.

        Now add in the fact that the average car on the road, in the US, is 11.5 years old (in 2015), and you have an environment which is extremely demanding.

        I personally would find it really interesting to see just how luxury car maintenance, injury and death statistics are between the Tesla and its actual peers: the BMW 7 series, Audi 8s, Mercedes 500s, and so forth. The cola can cars contribute a large percentage of fatality numbers so comparing a $60K or more vehicle vs. these is simply ridiculous.

    4. Eddy Ito

      The whole reason the OEM auto companies have been so much slower than Tesla on many of these technologies is that they have a

      slow plodding process that has only recently started to pick up pace and have been hamstrung by ridiculous regulations written by lawyers who don't understand how vehicles work. As example of the slow plodding I give you the humble audio cassette tape which pretty much was dead and replaced by CDs in 2001 when music companies dropped it but the auto industry kept it around largely as standard equipment for another ten years. For the regulatory silliness I offer up the sealed beam headlamp which was mandatory in the US for an absolutely absurd amount of time.

      Now of course things have changed and the car makers will adopt any insecure, internet connected, gee whiz IoT morceau de merde and stick it in without regard to the consumer. But that's ok, self driving or driven by some cracker on the internet, what's the difference really?

    5. 404

      I assert that the legacy automotive companies have generations of legal experience rather than engineering expertise.

      I submit as exhibit A: over a decade of accidents and deaths due to GM's faulty ignition switch. For a penny detent plunger that was millimeters too short...

    6. Black Betty

      Biggest problem with driver "assist" systems is human complacency.

      The in your face extra brake light initially reduced rear end collisions by 40 odd percent, but eventually settled at 4%, still worthwhile, but a poor comparison to the initial results. ABS braking was huge until the yahoo in us, figured out how much closer to the line it let us skate. US airbags are overpowered because they are designed to stop an unrestrained occupant. High visibility clothing somehow goes unseen.

      To the degree that they work it's all too often despite the driver.

    7. BillG
      Alert

      OEM Safety

      @captin_solo wrote: The whole reason the OEM auto companies have been so much slower than Tesla on many of these technologies is that they have a century of experience integrating technology into vehicles and testing and developing them over generations of model designs to ensure safe and reliable operation as well as recognizing that if they put tech into a vehicle, it could be on the road for 20 years. A time scale Silicon Valley never has to think about.

      Having ten years experience working in Detroit designing automotive electronics systems, let me back up everything Captain Solo has said.

      The OEM auto companies have a background of mistakes, failures, and lawsuits to back up their designs. Learn from mistakes. Learn from other's failures. In the beginning I'd be looking over some engineer's shoulder, looking at his schematic, and I'd ask something like "what is that capacitor doing there?" The answer was that, while not immediately necessary, there was a 1 in 5,000,000 chance that this and that would happen under less than ideal conditions, causing the car to stall/stop/shudder/etc, and that cap prevented that.

      Real world example: "Why are you grounding the heater core? There's no reason for it!", Well, when you have hot fluid going through a metal container a slight electric charge builds up, and over time (10+ years) degrades the core, so that eventually the core springs a leak. Not immediately obvious. There are hundreds of examples like that, and all that knowledge is in the most senior engineers (50yrs+) at Ford/GM/Chrysler/BMW who are cared for and coddled because they are the Keeper of the Secrets. The system allows them to mentor the younger engineers so they know the secrets.

      For me it was an eye-opening experience when I started working in Detroit. I came from computers and MIL-SPEC, and automotive quality control put them all to shame. Most commercial quality semiconductors were guaranteed to fail in cars within five years. Automotive Q-100 quality put MIL-SPEC to shame. Extremes of searing heat and chilling cold, constantly on, voltage spikes, extremes of dryness and humidity. Silicon Valley does not think like that. It's all get it out fast, release fast. If it fails three years from now, who cares, our next gen will be out by then.

      I can tell you there is an unofficial, informal sharing of safety-related information in the automotive electronics community despite the competitive nature of it all. If Tesla is not sharing, then they are not part of that safety community. Claiming something is an auto-pilot when the manufacturer of the system strongly objects to the term is putting sales over safety, something modern OEMs would never do and something the automotive community strongly condemns.

  2. bazza Silver badge

    Dangerous Trend

    "Unfortunately, with a few notable exceptions, Tesla also seems to be receiving Silicon Valley-style unquestioning coverage."

    And therein lies a considerable danger. Society, and by extension politicians, can easily be talked into a major change in the usage of an important piece of national infrastructure (the roads) by such unyieldingly favourable press coverage, simply because there's votes at stake.

    The real problems may come if the robust yet permissive attitude pioneered by California becomes diluted by weak politicians caving into popular demand. However, "popular" is not the same as "correct". Richard Feynman's sage comment about public relations and the unfoolable reality of nature applies.

    I applaud the robust attitude taken by the State of California - allow limited, supervised and controlled experiments with instrumentation and published results. Unfortunately for Google the results weren't flattering, and they clearly didn't like the results being published. Now we hear reports of discord within Google's team. That suggests to me that they've run out of ideas on how to improve their system, meaning they're going nowhere fast.

    Tesla's approach seems doomed to fail at this point (if self driving is what they're aiming at). Anyway, it's relatively pointless given that their major theme is "electric car". An electric car doesn't have to be self driving or auto piloted in anyway. Autopilot is a needless distraction from their real end goal.

    So if we want to progress towards provably safer self driving cars without the mass public experiment, that's going to mean a significant changes to what a road is. The only level of system we can prove to be correct using today's technology is essentially nothing more complicated than a unmanned railway (such as the Docklands railway in London, though even that has a manned control centre). If we made the roads more like railways then we could prove in advance that it would be better than today's human driven cars.

    However, that would mean no more motorbikes, cycles, pedestrians, road workers, deer, horses, other wildlife, pot holes, pedestrian crossings, snow, ice, fog, heavy rain, earthquakes, land slides, fallen trees, etc. on our highways.

    Can't see that happening.

  3. Anonymous Coward
    Anonymous Coward

    Tesla does have a weird reputation...

    And they're not exactly friendly in some ways either. First to think about the infamous Top Gear episode which made Tesla actually sue Top Gear not once, but several times. And they lost every time. I can understand that Tesla didn't like that episode, but really: what did they expect? The original Top Gear crew never made it a secret that they weren't exactly fond of electric cars. And it's not as if they hadn't done something similar before (they tried to take a road trip using electric cars, and ended up stranded several times because some cities didn't have any charging options).

    But I also remember the recent crash in the Netherlands. Tesla was plain out disrespectful with the way they handled the aftermath and cared more about their precious reputation than the family of the deceased. The problem: while the Dutch police were still investigating the crash and hadn't officially informed the next of kin yet Tesla felt the need to share a press release in which they shared a lot of details about the crash before the official police report came out.

    Tesla could have waited a week, or heck: 3 - 4 days or so, so that the family didn't have to hear all details from a press release, but apparently they obviously didn't care all that much about showing a bit of respect and courtesy towards those family members who had lost someone.

    And with all that in mind I wouldn't be surprised one single bit if Tesla did bend the truth a little here as well.

    1. Anonymous Coward
      Anonymous Coward

      Re: Tesla does have a weird reputation...

      > the infamous Top Gear episode which made Tesla actually sue Top Gear not once, but several times. And they lost every time.

      Just to say, you do not necessarily have to win a lawsuit for it to achieve its objective. In many cases, you do not even need to actually file one, just the threat of doing so is enough.

    2. MrXavia

      Re: Tesla does have a weird reputation...

      I thought the Tesla review was brilliant, and accurate...

      The problem for Tesla was they are an American company, so probably had never really watched Top Gear, and didn't realise that the test was to throw the car around the track and dangerous speeds by a top racing driver.

      If they had, they would never have let them near the car!

      1. Steve Todd

        Re: Tesla does have a weird reputation...

        The problem with the Topgear review was that it didn't represent what happened during filming. I've no problem with them saying it caught fire/ran out of charge etc if that's what actually happened on the day, but to misrepresent events isn't on.

        I can understand, in part at least, why they don't like electric cars (the G-Wiz is wretched for example) but Tesla are pushing them to the point they compete fairly well with Petrol (usable range, good acceleration and handling, not a track car), and that should have been more accurately portraid.

  4. Anonymous Coward
    Anonymous Coward

    Not Just Musk

    It's not just Musk that attacks anything seen as criticism of Tesla.

    Read the comments on any news story about Tesla, and you'll see commenters posting post after post replying/attacking any other posts they think are "anti-Tesla" I guess.

    Criticism just cannot be tolerated.

    Perhaps not unexpected, but I do wonder why all this emotional investment in what are essentially engineering issues.

  5. This post has been deleted by its author

  6. Herby

    It is all in a name...

    If you call something "auto pilot" expect the idiot behind the wheel to think it can do everything. If you call it "driver assist", there is a lower expectation of what it able to do (even if it does the same thing). In that case people won't doze off and end up trying to solve the problem of two things occupying the same space at once (which doesn't work well!).

    If you want a history lesson, look at the naming of MRI machines (Magnetic resonance imaging). Originally they were called NMR (Nuclear magnetic resonance) machines, but if you say "nuclear" everybody runs for the hills.

    Yes, it is all in the name! Why? Because nobody reads the manual and they never will!!

    1. Raphael

      Re: It is all in a name...

      Exactly this.

      Shortly after the crash that killed the chap, I was having a conversation with an older gentleman, and he quite logically said, "if it's an autopilot, why would I bother to have my hands on the steering wheel and pay attention to the road, an autopilot is meant to handle all that"

      And that's the point, calling it Autopilot means that people expect it to behave as an autopilot (regardless of what the manual says).

      1. Anonymous Coward
        Anonymous Coward

        Re: It is all in a name...

        Etymologically speaking, it *is* an auto-pilot, since it drives by reference to external features (the markings on the road, other vehicles when it can actually detect them).

        It's just not very good at it.

        1. tom dial Silver badge

          Re: It is all in a name...

          An autopilot would not have to be all that good to be better (and safer) than a great many human drivers I have observed over 50 years or so. Come to think of it, I can recall some occasions when I have to admit that a relatively simple machine might have driven more safely than I.

      2. Anonymous Coward
        Anonymous Coward

        Re: It is all in a name...

        Aviation autopilots don't manouver to avoid crashing into other traffic, so why should we expect automobile autopilots to do that?

        I'm not too familiar with nautical autopilots, but I would be surprised if they were able to manouver to avoid collisions.

        1. Dan 55 Silver badge

          Re: It is all in a name...

          Autopilots can land planes and most people who don't know about how autopilots actually work expect a similar amount of magic with a car's driver assist if it's called an autopilot. The first thing that has to change is the name.

          1. MrXavia

            Re: It is all in a name...

            Autopilots are powered by pixie dust, made from ground up pixies....

          2. imanidiot Silver badge

            Re: It is all in a name...

            As automating tasks go, landing a plane is pretty straightforward. Plane autopilots really are as intelligent as a pile of bricks. Yes they can fly the plane, but that's the easy bit of piloting. Any yahoo can learn to control a fixed wing. Doing so safely in busy airspace while maintaining positional awareness relative to other traffic is a whole different ballgame. And we are a long way from automating that aspect. We just work around it with minimum separation rules and constant ATC attention and guidance.

            1. Chz

              Re: It is all in a name...

              As someone else has said before, in the air it's extraordinarily unlikely that you're in a situation where ignoring the autopilot for 30 seconds will result in your decapitation. In a car, however...

          3. ~chrisw

            Re: It is all in a name...

            Autopilots are separate systems though more modern craft have control systems capable of controlling the plane in cruise, approach, landing and rollout. Takeoff and taxi needs a human. In flight, autopilot typically maintains a bearing and altitude whilst keeping the plane level; it has pitch, yaw and angle sensors plus accelerometers and gyros (inertial guidance) to understand its relative speed and axes of travel (and perhaps GPS and DME).

            Then there's autoland, where the plane follows an ILS or MLS beacon in combination with a radar altimeter for near-ground final manoeuvres. I personally think, whilst marvellously precise systems (albeit with design-controlled error rates) they're comparatively dumb systems. And that's good, because what can you do if a really complex system fails at 10,000 feet in thick fog mid landing approach?

            Airspace is so intentionally precisely controlled it allows for variation in craft behaviour -- much more than a car can get away with. Plus you have, at minimum, three small armies of highly skilled human operators watching over flightpaths at every takeoff, cruise and landing, advising the two highly skilled human operators in the cockpit.

            Current car 'autopilot' systems involve visual spectrum cameras and radar, GPS, collision detection and avoidance. The better ones also perform real-time trajectory projection (to figure out if something that's moving unpredictably might intersect with you given your current rate and direction of travel). On top of that, they can handle proximity warnings, can read signage and lane markers and - in Tesla's case at least - machine learning as their fleet progressively builds up a database of road conditions and optimum speeds. Those systems are advising and watching out on behalf of one (probably average-skill, and likely distracted) driver.

            Autopilot systems haven't really changed that much in decades because of all the failsafe precautions put in place - and the controlled operating environment. You stick a car on the road, who knows what will happen in 30 seconds' time... I think the current generation of systems is already borderline performing magic tricks. Though I wouldn't quite trust it yet with my commute. Humans are crazy.

        2. Cynic_999

          Re: It is all in a name...

          "

          I'm not too familiar with nautical autopilots, but I would be surprised if they were able to manouver to avoid collisions.

          "

          Some can, providing the conflicting object is a vessel that has a functioning AIS system. A marine autopilot is however usually quite happy to steam straight into a harbour wall at full tilt, and it's not unheard of for a large ship to arrive in port with bits of a sailboat hanging from its bow anchors ...

    2. SJA

      What about Drive Pilot

      Is that a better name?

      1. werdsmith Silver badge

        Re: What about Drive Pilot

        There are enough warnings and instructions thrown at the driver of a Tesla who uses Driver Assist, ignoring them has to be a willful act. I've never driven a Tesla and it is plain obvious to me that Driver Assist has to be closely watched.

        It's getting all a bit Stella Liebeck, and we are all having to pay for business overheads because they have to cope with the lowest common denominator.

        1. SJA

          Re: What about Drive Pilot

          Well, I was mentioning Drive Pilot because that's what Merc calls it's "Autopilot"...

      2. DrM

        Re: What about Drive Pilot

        How about Cruise Control+ ?

    3. Jonathan Richards 1

      Re: It is all in a name...

      ... indeed, and not just the name 'autopilot'. There is a strong feeling in many quarters that Nicola Tesla himself was a wronged genius, done down by The Powers That Be, etc. etc. I sometimes think that the Tesla brand operated by Mr Musk feeds off that association.

  7. Anonymous Coward
    Anonymous Coward

    Looks like Musk is trying to grow an ISIS beard in the lead photo. Just saying....

    1. Sorry that handle is already taken. Silver badge
      Joke

      ISIS beard?!

      That looks more like a neckbeard, which seems appropriate.

  8. Fan of Mr. Obvious

    We are not liable for...

    Tesla is acting like a typical software company. I have commented several times in the past that my personal opinion is autonomous vehicles are going to run into problems because the manufactures are not going to want to accept liability -- they want to act like software companies instead of vehicle manufactures. Given that the autonomous vehicle community is spinning the tech as something that will save consumers live, and money on insurance, I am guessing they have no intentions of every accepting liability -- the autonomous car is/will be perfect in their eyes so no need to look to them for problems.

    If Telsa accepts liability for anything accident related, they open the door a little. Every time the door is inched open an attorney, or group of attorneys, will poke their heads in and say hello. They likely prefer the strategy of keeping the door locked at the moment - can't say I blame them.

    Still, much work must be done in regards to changing liability laws for companies to be off the hook. If they do succeed, it will be you and I to blame for everything. We used the wrong soap when we washed the car therefore it left a film on the sensor; the tires were not balanced properly therefore... And we thought dealing with vehicle warranty issues were bad.

    1. Sorry that handle is already taken. Silver badge

      Re: We are not liable for...

      Volvo intends to accept liability for what its AVs do.

      1. Fan of Mr. Obvious

        Re: We are not liable for...

        Indeed, Volvo did say that. However they also said the US must work out laws regarding liability first -- the rules are just not clear enough for Volvo.

      2. JeffyPoooh
        Pint

        Re: We are not liable for...

        Volvo...king of serial Self-Braking failures.

        Two separate incidents on YouTube.

        All these folks need to tune down their hubris.

        "A.I. is hard. Especially in the real world."

    2. Neil Barnes Silver badge

      Re: We are not liable for...

      Nonetheless, one might not be terribly happy getting into a car which comes with the sort of licence most software seems to: "This XXX is not guaranteed to operate correctly or even at all..."

      It seems to me that cars are suffering the same feeping creaturism that software products do; no-one actually *uses* a particular feature, but it ticks a box for a reviewer, so we'll stick it in. And then another, and another...

      What you need in a car is some suspension, steering, engine, and a chassis/body. All mature technologies which can be implemented without *any* electronics (I ignore engine control systems for the sake of argument; they're not essential except by legislation). But then someone comes along and things a cruise control would be nice, and someone else thinks it would be nice if it matched speed to the car in front, and then a system that knows the speed limit, and then one that can take avoiding action if the car in front does something silly, and then, and then, and then... before you know it you've somehow acquired an autopilot - which may or may not work quite as intended.

      Tesla needs to decide whether it's in the car making business or the autopilot business. They're not the same, and I hope they decide for the former - electric cars with a decent range are an excellent idea if only because the pollution they cause is probably easier to deal with in bulk. But if I don't want to drive myself, there's still taxis, trains, buses, professional chauffeurs... I don't think that the autopilot is finished until it can do a drive through a city centre at rush hour *and* a trip through the mountains. Until then, hands on the wheel please.

  9. Anonymous Coward
    Anonymous Coward

    Mobileye

    If you visit the Mobileye website and look at their technology solution it's quite clear that they are rooted in the driver assistance market and have no chance of ever extending it to allow autonomous capability. As such it is not surprising that they are playing the safety card as technology leapfrogs them, first relegating them to being a second layer of safety monitoring "autopilots" before advancing to the point where Mobileye's mono-vision technology ceases to give tangible benefit.

    The question is are they right to claim autopilots aren't safe because "drivers" will ignore the road and trust them despite them not being up to the job? This really depends on whether the autopilot is designed to drive at the speed the human driver could or it slows down to a speed consistent with how the sensors and processing system cope with the prevailing conditions. If this is too slow for the human they can disengage the autopilot and go faster. This is temporary: the technology will get faster.

    1. Marcus Fil

      Re: Mobileye

      Yes, but... we possilbly could have an "autopilot" complete with computer vision, LIDAR etc. that could safely negotiate the centre of Manchester at rush-hour at a top speed of 7 mph. However, at some point it is going out onto faster roads and if it is still going 7 mph it is going to get stopped by the cops (has it even been programmed for that?). Speeding up the processing, LIDAR sweep rate, adding more pixels to the camera etc. still does not solve the fundament problem of dealing with the unexpected at sensible, usable road speeds. Why? Because it is hard to program for the unexpected. Mobileye sensibly say "driver assist " because their technology is an aid, not an abdication of resposiblity. Are they being pragmatic in recognising that technology cannot replace a little light thinking? Those who advocate "autopilots" as the answer to safe driving for all need to spend time awy from their simulators and keyboards and much more time behind the wheel - all around the clock, all around the world and in every climate and season. Once you appreciate the scale of the problem see if you remain so confident "autopilots" can fix it.

  10. a_yank_lurker

    Autonomous Cars

    Autonomous cars are not ready for the public at this point. They are purely engineering test beds. What many automakers are doing is incorporating bits they believe will help improve safety such as automatically braking under certain situations. A couple of reasons for this: autopilot systems are still too buggy for the public and public is not quite ready for them. Pretty much everyone who is working on an autonomous car has said that the technology is not ready but we are "close". It may be like nuclear fusion which is always "a few years" in the future for 50 years+.

    One of the problems with an autonomous vehicle is that vehicles can have a useful life of 15 years+ with proper maintenance, a little luck, and moderate use. Owners will expect similar results with an autonomous car. Many advertise that many owners get this kind of life so expectation is there. So the computer systems must work for the entire mechanical life of the vehicle which often will be 20 years.

    Too name a driver assist package "Autopilot" is marketing stupidity at best if not risking a regulatory investigation for false claims.

    1. JeffyPoooh
      Pint

      Re: Autonomous Cars

      "...autopilot systems are still too buggy..."

      I'm not sure that quite captures the gulf between today's pathetic attempts and what the actual solution will look like so many years from now.

      What you've written is like stating that fusion power is still just a bit immature. Or flying cars aren't quite ready for prime time. Or peace in the Middle East isn't expected in the coming months.

      1. bazza Silver badge

        Re: Autonomous Cars

        @JeffyPoooh,

        "I'm not sure that quite captures the gulf between today's pathetic attempts and what the actual solution will look like so many years from now.

        What you've written is like stating that fusion power is still just a bit immature. Or flying cars aren't quite ready for prime time. Or peace in the Middle East isn't expected in the coming months."

        Quite. And it's a gulf the size of which the stats published by California unequivocally illustrated, to Google's apparent irritation.

        You have to feel for the dev team, to have the true scale of their task writ large for them to fully comprehend. Yet it was always a challenge the scale of which was easy to comprehend with scholarly study of things like airliner automation, automated train systems, the first Ariane 5, etc. All these things point to it being a really tough challenge, and only one of them is actually fully automated and carries people (and operates in an artificially simplified environment; a railway). Given that there hasn't been a fundamental break through in machine comprehension (the sort that is demonstrably infallible, not just slightly better than the last one in lab demos), any old fool could conclude that a fully dependable self driving car isn't a viable option at this time.

        You'd have to be really young and cocky to think that you could do better than all those predecessors. Well, Google are a young cocky company with, apparently, a lot of young cocky engineers. If they want to learn the lessons of life the hard and expensive way, so be it, that's up to them and their investors (a fool and his money will soon be parted. Anyone got shares in Tesla, Uber, etc?). So long as they don't manage to con everyone else into believing it works and get such things mandated by law...

        The rest of us who have been there before will puff contentedly on our pipes whilst sitting in our most comfortable armchair and tells tales of yore...

        1. bazza Silver badge

          Re: Autonomous Cars

          One other thing. The engineer who tells his company that something wildly ambitious cannot be achieved is worth a lot of money. They may very well be saving them from spending a fortune for no net gain. Certainly their point of view should be represented to the shareholders before the board decides to overrule them.

          I suppose companies like Google and Apple, and maybe Tesla and Uber, have got money to burn on experiments such as self driving cars, but it's still their investors money, not their own personal funding to do with as they wish.

          1. Grunchy Silver badge

            Re: Autonomous Cars

            I think Tesla is embracing self-driving tech because focusing solely on an all-electric powertrain is not all that remarkable, and very easy to replicate by competitors. I'd buy a Leaf or a Volt or any other thing (hybrid) over any Tesla, if that was all they had going for them.

            Tesla cannot afford to not have a working self-driving technology to differentiate them from other car companies. They have never turned a profit, and they may never, now (and especially with so many other companies now focusing on creating their own flavor of that technology). The actual car companies are going to run Tesla out of business if they don't bring in new deliverables such as a whole-home battery and a home-generating solution.

            I won't miss Tesla when they're gone. I do miss Aptera though, that was a sweet looking machine...

    2. L05ER

      Re: Autonomous Cars

      to claim that autopilot is at all misleading is stupidity at best.

      autopilot doesn't do what you think it does, and that is not tesla's fault.

      the name is very accurate... and i'd argue better than its namesake. unless of course plane autopilot has any obstacle detection and avoidance capabilities i'm unaware of...

      1. Anonymous Coward
        Anonymous Coward

        Re: Autonomous Cars

        I think the stupidity is all yours.

  11. Mark Solaris

    Tesla is one of the automotive companies that are actually developing the technology. Developing means adjusting and refining and enhancing. One of those enhancements is to step away from a part manufacturer that isn't on board with their direction. Another enhancement, due to arrive imminently, is to add heavier restrictions to the Autopilot, insisting on the drivers hands being on the wheel much more, even going as far as requiring the car to be stopped and in Park before it'll allow the Autopilot to be re-engaged. This will catch the idiots that don't read manuals, are too arrogant to follow the rules and then rush for a lawyer when they cause an accident from using the vehicle the wrong way.

    1. Nate Amsden

      i think most people

      Would prefer to have them develop the technology in private(like google or others) or at least slap a big BETA sign on their cars. So people that do want to play beta have a chance to, but most of the rest that don't understand will be better informed to look elsewhere.

      1. sleepy

        Re: i think most people

        You can't enable the Tesla autopilot without first explicitly acknowledging that it is driver assistance beta software, and that it must be constantly supervised. And you can't engage it subsequently without keeping hands on the wheel. The enforcement of hands on wheel may be a bit lax, but the car will not stay in autopilot without hands on the wheel.

        1. Anonymous Coward
          Anonymous Coward

          Re: i think most people

          The enforcement of hands on the wheel isn't "lax", it is practically non-existent. It won't disengage until you ignore three warnings to put your hands back on the wheel over the course of an hour. So basically once an hour you have to briefly place your hands on the wheel when it warns.

          IMHO it shouldn't allow you to take your hands off the wheel AT ALL for more than 10 seconds without disengaging. If you can take your hands off for minutes at a time, what's the incentive to be paying attention? It encourages people to text, check email, rummage around in the back seat to find something, etc. Tesla can put up all the disclaimers discouraging such things all they want, but the lack of enforcement in the software shows they're basically winking at you and saying "do what you want, we won't stop you, we're just telling you not to because our lawyers said we should".

          1. mark 177

            Re: i think most people

            You have to hold the wheel once an hour? It depends on conditions, but it's actually more like a minute or two. And I own one, so I'm not making this up.

  12. L05ER

    all these implications...

    where do you get them?

    the only message i ever got from the "blame shifting" as it was called... was that the system as a whole was (as designed at the time), insufficient to handle the circumstances it found itself in. period.

    none of the BS this article seems eager to shove down people's throats... but then what should i expect from a publication that has been nothing but hostile toward the autopilot system. everything i know about its namesake and everything the company itself has stated is far and away from how it is presented here.

    do you people complain about traction control because it doesn't maintain traction 100% of the time without fail? ABS because they still momentarily lock in very wet conditions?

  13. werdsmith Silver badge

    the only message i ever got from the "blame shifting" as it was called... was that the system as a whole was (as designed at the time), insufficient to handle the circumstances it found itself in. period.

    Those circumstances are mostly a driver ignoring all the warnings in the manual, on the big screen in the car and from a human briefing when it is purchased. Quote:

    "Warning: Traffic-Aware Cruise Control has limited deceleration ability and may be unable to apply enough braking to avoid a collision if a vehicle in front slows suddenly, or if a vehicle enters your driving lane in front of you. Never depend on Traffic-Aware Cruise Control to slow down the vehicle enough to prevent a collision. Always keep your eyes on the road when driving and be prepared to take corrective action as needed. Depending on Traffic-Aware Cruise Control to slow the vehicle down enough to prevent a collision can result in serious injury or death.

    That comes from two pages of similar warnings.

    1. Cris E

      There's a troubling gap between calling something autopilot in the ads and Traffic-Aware Cruise Control in the disclaimers. Either the thing is a magic unicorn that lets you text, watch movies, have sex, etc while driving or it's only cruise control with some new bells and whistles. But Tesla is selling it one way while simultaneously trying to limit responsibility and that's not right.

      Mobileeye is certainly uncomfortable with the difference, to the point where they're getting out of one of the best relationships their company could possibly hope to be in. It's telling that they fear this liability more than the potential profits they could realize by staying with Tesla. Even if this turns out to be a You Can't Fire Me I Quit moment they still aren't letting this be sold as autopilot.

  14. dloughlin

    I own a Tesla Model S and use autopilot daily. It's perfectly safe, very relaxing and a great safety feature.

    Tesla make it very clear you must keep your hands on the wheel, right from the hand over when you pick up the car, to the warning/T&Cs when you enable autopilot, which is disabled by default, and in the manual. It also reminds you every time you activate autopilot, and regularly while driving.

    Brown ignored all this and was supposedly watching a movie. His crash was human error. Just the same as it would be my fault if I crashed driving 150mph on a windy country road (something that happens daily, yet doesn't make the news) - I used the car inappropriately.

    Tesla are dragging the auto industry into the future. Some people don't like change. And most companies don't like competition.

  15. Northcroft

    The article lacks basic research

    The author mentions Joshua Brown, the only real fatality so far in the USA. A quick browse of the internet reveals that Joshua Brown made several YouTube videos about his Tesla. He was very proud of it. In one of the videos makes it clear that he knew about the problem that killed him!

    Joshua was killed in Florida. I think he lived somewhere near the Canada border. So he set his car speed 10 mph faster than the speed limit, and starting watching a video. Did he have an 18 hour drive ahead of him? Has anybody asked questions of the lorry driver who pulled across the road? Could he have avoided the accident, in spite of the 10 mile an hour excess speed of the Tesla? A Google map of the incident shows that the road was very straight.

    Of course any idiot could have killed himself driving a Tesla with autopilot. And 1,500,000 do every year, without autopilot.

    I think the death toll is now two - a chinese man drove his Tesla into a back of a lorry parked in the fast lane of a motorway ... we do not know if autopilot was turned on. What was the lorry doing parked in the fast lane?

    The accidents are the price we pay to learn how to make things work. The car industry did not fit seat belts for the first 75 years ... how many died? The USA still does not make lorries of the kind that killed Joshua Brown have barriers on the side of the lorry to prevent cars going under them. A friend of mine was killed in the UK 30 years ago in the same way - an accident that cannot happen any longer as European law requires all such lorries to have side skirts - or whatever they are called. And they save the lorry driver money, as fuel consumption is reduced ...

    The Israeli company making the sensors and chips that the Tesla uses are focussing on their existing clients. Obviously. They have not got the balls of Elon Musk ... fair enough. But somebody has to figure out how to do it, and Elon Musks method is very practical - and he accepts that some lives have to be lost - but he is making the very important point that it is the balance that counts. Lives saved against lives lost. He is doing it in exactly the right way - starting with building the safest car ever built - and experimenting with that car ... doing everything possible to minimise the risks. How many lives has that car saved ... versus how many have died. I think the balance is heavily in his favour, and is only going to increase - the factor of 10 that he is aiming for cannot be far away.

    How bad is that?

    1. Test Man

      Re: The article lacks basic research

      " and he accepts that some lives have to be lost"

      Don't be so fucking stupid.

  16. Steve Davies 3 Silver badge

    Musk is a Salesman.

    and purveyor of Snake Oil.

    Do I really need to say anything else?

    1. James Hughes 1

      Re: Musk is a Salesman.

      You could say you are talking out of your arse perhaps?

      Snake oil salesmen sell snake oil, not 50k cars a year ramping to considerably more, or successfully reduce the cost of a orbital launch by 50%. Or found Paypal for that matter, which I believe is a very successful company.

      I've never understood this style of denigration of Musk. He's clearly extremely successful at what he does, everyone who has met his says he's one of the smartest people they have ever met, and he is single handedly (almost) trying to revolutionise Space, solar and cars, to the benefit of all.

      1. Anonymous Coward
        Anonymous Coward

        Re: Musk is a Salesman.

        Two people in a balloon. One called Eben, one called Elon.

        One of them has to go.

        It's an easy decision, isn't it, especially when you understand the backstory behind Musk's approach to "designing" reusable rocketry (which can be summarised only slightly unfairly as "never mind the analysis, let's build something and see how well it works" - rather like his Autopilot, perhaps?).

      2. Steve Davies 3 Silver badge

        Re: Musk is a Salesman.

        Sigh...

        Snake Oil salesmen were purveuors of magic potions that promised to cure every disease known to man.

        I don't trust him simply because he has far too many fingers in too many pies for him to concentrate on making Tesla the best. Sure they are innovating but he keeps on making promises he can't keep.

        Calling the 'driver assistant', Autopilot was a mistake. Anyone involved with Aircraft autopilots will tell you that.

        I don't see any evidence that the software has been developed and tested to the same standards as Flight Critical software. If I ever drove a Tesla, I would not engage Autopilot until it gets a lot better and the code is certified as being tested to Flight Critical standards. That alone would defer an awful lot of lawsuits.

        So in my opinion, he is a snake oil salesman.

        Naturally I could be wrong but so far I see lots of promises from Tesla and lots of money being lost.

        When the model 3 is delivered for £30K on time I might revise my opinion of him.

  17. Spudley

    To be honest, I'm not really that interested in the self-driving or driver assist technology. I've got adaptive cruise control in my own car, and I've hardly ever used it. If I had a Tesla, I don't expect I'd make any meaningful use of 'Autopilot' either.

    What interests me about Tesla is the powertrain -- they've got an electric car with double the range of the next best available vehicle, and half the time to charge it. But more importantly, they've spurred some of the rest of the industry into a response.

    I would love for my next car to be fully electric, but it will have to be in my budget and good enough to justify over a petrol engine. Tesla's Model 3 might just hit the spot. But more importantly, the competition that's being driven by the Model 3 means that there are a number of other cars that should also be available in the next year or two that might be an even better fit for me. I'm looking forward to them being available, and I seriously doubt that they would be on the market anything like as soon if Tesla wasn't pushing the industry.

    The self-drive stuff... that's all well and good, but the real value of Tesla is in the batteries.

  18. Wils

    Self driving bullshit

    What can people expect from the folly that is Tesla, produced by Musk and his madmen, bought by rich, technology fan-boy's.

    Return full control to the driver, putting a huge spike in the centre of the steering wheel is the only technology needed to improve safety. Self driving is total fantasy, collision avoidance systems dangerous. How long will this utter madness go on being promoted.

    1. Anonymous Coward
      Anonymous Coward

      Re: Self driving bullshit

      Quote

      What can people expect from the folly that is Tesla, produced by Musk and his madmen, bought by rich, technology fan-boy's.

      Could Tesla be the new home for the former disciples of St Jobs?

      After all Apple is so boring these days.

  19. Anonymous Coward
    Anonymous Coward

    Brown watching a video...

    Is bull. The Florida police have already indicated there was no evidence for this. There was a portable DVD in the car but no evidence it was running at the time. Also, the eyewitness account for this is the driver of the truck that was driven into, so until the official investigation concludes we've got to take his version of events (like all others) with a pinch of salt, as we don't know if there could be a degree of blame shifting going on. There is a suggestion the truck driver could have heard an audiobook playing, and assumed the driver had been watching the movie "on that big screen", but again it's all speculation.

    If you look at the crash location on Google Maps, it's quite an awkward turning for a truck. Get it slightly off and a truck drivers view of oncoming cars could easily be compromised. Given that it's otherwise a wide, straight highway, I think there's more chance it happened because a truck tried to complete a manoeuvre too late, or misjudged the speed of an oncoming car, or had difficulty seeing it in the time available. Total speculation of course. But for the car to have hit it where it did, the truck had to have been performing a late maneouver, or just doing it much slower than was safe. If something drives out right in front of you, and you're going at highway speeds or above, all systems human or otherwise are going to struggle to avoid it. Though slamming on the brakes might have at least made any impact more survivable.

    One would hope that successive software updates such as 8.0 which allegedly deal with scenarios exposed by this accident reduce the likelihood of that kind of accident ever happening again. That has to be the holy grail of all river assistance systems, actual reductions in fatal accidents through recognising scenarios and avoiding them. As someone who briefly fell asleep at the wheel for a fraction of a second the other week, and yes I'd been taking breaks, just obviously not enough for such a dull road, this new world can't come soon enough.

  20. Mark 85

    Never underestimate the stupidty of humans...

    ... which seems to be the problem. There's the urban legend about the two guys in an RV (motorhome/caravan) driving down the freeway and the one is in the galley making a sandwich. The other shows up and starts to make one. The first says..."why aren't you driving?". The response: "I've set the cruise control and it seems to be controlling just fine." This moments before the RV ran off the road.

    Yeah.. legend. But there's a lot of people that stupid. Driving and Texting? Driving under the Influence? Those are two biggies. There's a million other things that can and will go wrong. Calling something "autopilot" ranks right up there especially when it isn't. It's a glorified cruise control with features...

  21. DrM
    Mushroom

    Not just me?

    I've never believed a word he has to say. Well written.

  22. tr1ck5t3r

    So these cars have driven 100k miles on average and handed back to the driver to avoid an accident every other mile or so, but dont dwell on the last part its not good for sales.

    Give me an autonomous vehicle that can complete a journey _without_ handing back to the driver and I'll call it autonomous.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like