The Unz Review - Mobile
A Collection of Interesting, Important, and Controversial Perspectives Largely Excluded from the American Mainstream Media
 TeasersiSteve Blog
Do 737 MAX Crashes Bode Poorly for Self-Driving Cars?
🔊 Listen RSS
Email This Page to Someone

 Remember My Information


Bookmark Toggle AllToCAdd to LibraryRemove from Library • BShow CommentNext New CommentNext New ReplyRead More
ReplyAgree/Disagree/Etc. More... This Commenter This Thread Hide Thread Display All Comments
These buttons register your public Agreement, Disagreement, Troll, or LOL with the selected comment. They are ONLY available to recent, frequent commenters who have saved their Name+Email using the 'Remember My Information' checkbox, and may also ONLY be used once per hour.
Ignore Commenter Follow Commenter
Search Text Case Sensitive  Exact Words  Include Comments
List of Bookmarks

iSteve commenter Jack D writes:

One of the weaknesses of automated systems (they are finding this also with self-driving cars) is that they are designed so that if something goes wrong that is beyond the design parameters of the system, they suddenly snap themselves off and return control to you (or in some cases, you snap them off because they are acting funky). The thing says “I give up. Something is wrong but I’m not sure what – human you figure it out and if it crashes it will be your fault and not mine.” Either way, suddenly you are back in control of the plane/car and there is something wrong to begin with that caused the automated system to flake out, but you have been kind of half-dozing and are not mentally prepared to figure out what is wrong AND in meantime pull the plane/car out of a precarious attitude.

In the case of Air France [crash in 2009] it would have been better for the pilot to have done nothing. The stuff that he did made the situation much worse. But that’s supposed to be the pilot’s job – it’s like being a fireman where 99% of the time you sit around and eat pizza (actually nowadays a lot of paid fireman double as EMTs) and 1% of the time you save people (including yourself) from certain death.

I’d like to have an automated system on my car for low-speed parallel parking. But for full speed stuff? What does it gain me if I have to alertly manage my automated car all the time so that I can suddenly take over and go all Captain Sullenberger?

Hide 200 CommentsLeave a Comment
Commenters to FollowEndorsed Only
Trim Comments?
  1. anon[375] • Disclaimer says:

    I’d like to have an automated system on my car for low-speed parallel parking.

    I am the king of parallel parking. It’s an art. Don’t take away my paintbrush.

    • LOL: Alfa158
    • Replies: @Jack D
    , @Prester John
  2. Anon7 says:

    Tesla’s Navigate on Autopilot does a pretty good job of doing the driving on highways, getting into the correct lane, taking the right exit, merging with traffic on the next expressway. However, it is assumed that you, the driver, are always paying attention to what the car is doing; the car is still your responsibility, so you need to take over if you’re not comfortable with what the car is doing. (These interventions are uploaded to Tesla for analysis and help improve the system.)

    What does it gain you? It’s way more relaxing to watch the car drive than to drive yourself. I think it’s because the element of competition with other drivers is completely removed. The goal is to follow directions and get to your destination safely at a set speed, rather than to pass that guy who cut you off at the last interchange.

    So far, a car that drives itself at speed on highways without the driver’s attention is still science fiction.

  3. Anon[321] • Disclaimer says:

    OT: It has just been revealed that Beto O’Rourke used to belong to a hacker group. Just when we get rid of Zuckerberg’s presidential aspirations, this prying nitwit shows up. Oh, and he used to write very stupid and rather gross poetry.

    • Replies: @El Dato
    , @IHTG
  4. Noting crashes of this nature are fine.

    But the hold up for self driving cars are the accidents involving self driving cars.

  5. Anonymous[427] • Disclaimer says:

    So far, a car that drives itself at speed on highways without the driver’s attention is still science fiction.

    Autonomous Dreaming Today’s hot-button autonomy discussion rages on, but in 1964, GM was fully invested in developing a car for “automated highways,” as it called them. The Firebird IV concept was that car. The Firebird IV debuted at the 1964 New York World’s Fair, and though claimed by GM to be powered using a turbine (it never actually ran or drove). The idea was a system of electronic highways that would allow the vehicle to run autonomously, using a system of radars to monitor progression. It was the final project in the Firebird series, and the most progressive of the four. It also featured a dramatic new design, and an interior loaded up with luxuries like a television, a game table, and even a fridge.

    The Firebird IV never ran but the Firebird III did and still does, on occasion:

    GM built the third design, the Firebird III, in 1958 and debuted it at Motorama in 1959. It is another extravagant concept with a fiberglass body and no fewer than seven short wings and tail fins (which were tested extensively in a wind tunnel). It is a two-seater powered by a 225 hp (168 kW) Whirlfire GT-305 gas turbine engine, with a two-cylinder 10 hp (7.5 kW) gasoline engine to run all the accessories. Its exterior design features a double-bubble canopy and technical advancements to make it more practical, such as cruise control, anti-lock brakes, and air conditioning. It also featured “Space-Age” innovations, such as special air drag brakes like those found on aircraft, which emerged from flat panels in the bodywork of the car to slow it from high speeds; an “ultra-sonic” key that signaled the doors to open; an automated guidance system to help avoid accidents; and “no hold” steering. The driver controlled the steering with a joystick positioned between the two seats.[7] This gave the car a more futuristic feel and simulated the experience of flying a plane.

    The General Motors Firebird comprises a quartet of prototype cars that General Motors engineered for the 1953, 1956, and 1959 Motorama auto shows. The cars’ designer, Harley Earl, took his inspiration from the innovations in fighter aircraft design at the time. General Motors never intended the cars for production, but rather to showcase the extremes in technology and design that the company was able to achieve. The cars recently joined the display at the Henry Ford Museum in Dearborn, Michigan, and still make regular car show appearances. The tradition of offering prototype vehicles continued with the Pontiac Banshee series.

    The automotive gas turbine was one of those things that almost got produced but never quite did. Chrysler in the US and Rover in England had the most successful engines, they actually were productionworthy, although Chrysler’s fourth gen engine (of which maybe sixty were made) had some limitations, like not having a decent way to run A/C with the car not in motion. In 1963, though, unless you lived in certain really hot and humid areas, that wasn’t the deal killer it would be now.

    • Replies: @Jack D
    , @Alfa158
    , @Philip Owen
  6. I think high speed freeway driving is the easy part that is best addressed by AI. Complications arise on local roads when dealing with pedestrians, buses and general idiots.

    Perhaps the best approach is dedicated freight freeways that run on newly built highways over railroad rights of way. Alternatively, there can be dedicated freight truck lanes on existing freeways that are separated from the human drivers. Once the trucks hit the local road, a driver should jump aboard.

    100% automated driving is coming, but it is at least a decade away.

  7. I’d like to have an automated system on my car for low-speed parallel parking.

    I want computer control to park like Elwood

    • Replies: @Hypnotoad666
  8. Hoff says:

    Luxury cars have been able to auto-parallel park for around 4 years now.

    I have an acquaintance who is a physical therapist who bought a Tesla. He goes to each of his clients homes. This takes him, throughout his day, from the Valley, to beach cities, to deep into Orange County. He said his job would be impossible without the self-driving ability of the Tesla, especially in stop-go traffic, which he must endure every day. In that scenario, it allows him to do “office work” while the car drives itself in slow heavy traffic. The car cost quite a bit more than he would have preferred to spend, but without it, he couldn’t handle the job.

    • Replies: @Hibernian
  9. Self-driving cars are like high-speed rail; usually ok, but the problems and expense are spectacular.

  10. jim jones says:

    I prefer to take Public Transport and let the bus driver do all the work.

  11. El Dato says:

    Another take

    We’re told that O’Rourke wasn’t involved in the more nefarious parts of the hacker ring’s activities, such as the development of tools to control hijacked Windows PCs. He did admit to circumventing phone bills and downloading cracked software, a common practice among members of underground BBS groups at the time.

    That O’Rourke would be involved with the legendary Texas hacking crew is surprising, but hardly out of character. Part of O’Rourke’s cachet in left-wing circles is his past as a punk rock musician in the US state, and strong counter-culture leanings.

    At least someone out there in large Outer Cubicle understands computers.

    • Replies: @Steve Sailer
    , @AnotherDad
  12. IHTG says:

  13. Realist says:

    Driverless cars are dangerous. If you don’t want to drive take public transportation.

    • Agree: jim jones
    • Replies: @Jack D
  14. @El Dato

    Wasn’t Mencius Moldbug in Dead Cow?

  15. LondonBob says:
    @Clifford Brown

    Dedicated freight freeways just sounds like trains again.

    Self driving will never happen, they haven’t even perfected it for planes where a lot more can be spent and there is less to monitor.

    • Replies: @Boswald Bollocksworth
  16. El Dato says:
    @Clifford Brown

    I think high speed freeway driving is the easy part that is best addressed by AI. Complications arise on local roads when dealing with pedestrians, buses and general idiots.

    The problem then occurs if there is road work on the freeway. Suddenly you are driving at speed into something that is mightily confusing to the neural network processor … raaah! “She canna’ take no moar, Captain” as Engineer Scotty used to say.

    Here is something on the Uber fatal road accident of 2018-03-18 where the Uber’s system, possibly assembled in a bit of haste, with less sensors than there should be and maybe with less than stellar engineering (why is Uber even in engineering?) ran over a pedestrian walking her bicycle over the street:

    Uber won’t face criminal charges after its robo-car killed woman crossing street

    “After a very thorough review of all the evidence presented, this Office has determined that there is no basis for criminal liability for the Uber corporation arising from this matter,” Yavapai County Attorney Sheila Sullivan Polk said in a letter to Maricopa County Attorney Bill Montgomery.


    (IIRC there were rumors that the pedestrian was stoned, but I don’t see how that has anything to do with the case at hand; sounds like info management to me; same as that YouTube video of the car’s POV that was a bit too dark than it needed to be *cough*).

    The National Transportation Safety Board (NTSB) is still looking into the incident but released a preliminary report last June. The report says Uber’s software kept reclassifying the pedestrian as it approached, first as an unknown object, then as a vehicle, then as a bicycle, which affected how the system plotted an obstacle’s expected travel path.

    It is then said that the Uber system could have applied emergency braking if it hadn’t been disabled because its propensity to emergency break was a bit too high than was comfortable for road tests. Aka. “The alarm sounded all the time, so we disabled it”, a standard … err, trope … of event sequences leading to suboptimal outcomes.

    “At 1.3 seconds before impact, the self-driving system determined that emergency braking was needed to mitigate a collision,” the NTSB says in its report summary. “According to Uber, emergency braking maneuvers are not enabled while the vehicle is under computer control to reduce the potential for erratic vehicle behavior.”

    Okay, so. Uber is held not criminally liable (I don’t know about any civil lawsuits in the pipe) but the second-in-command driver, aka the “safety driver” (rumored to sport a less than pristine criminal record, but again so what and where does that info even come from) is not in the clear:

    The case heads back to Maricopa County which now has to decide whether any charges against the safety driver, who was supposed to be monitoring the vehicle, might be warranted.

    This does not bode well for the future of the self-driving car. “You weren’t monitoring the self-driving vehicle, pay up!” pours cold water on the concept of the self-driving car itself. If you have to be alert at all times to “expect the unexpected”, you might as well drive yourself.

    Additionally, those neural networks will have to explain themselves and car log files will have to look a lot like Terminator POV videos (without the 1970s teletype printer sound effect).

    • Agree: Hibernian
    • Replies: @Jack D
    , @El Dato
  17. I use lane hold assist pretty much all the time, and for the most part it gives me the kind of ride one got in early F-111’s in terrain following mode. It’s difficult to trust for more than minor corrections.

    This is certainly going to delays the dreams of fully automated airliners, but since car crashes kill perhaps only one or two dozen, it will slow things down, but not very much, and it certainly will not kill them entirely; that will be done by the greenies.

    • Replies: @Old Prude
    , @Simply Simon
  18. Old Prude says:

    When you’ve dccline into surly decrepitude and your kids are trying to take your keys away, and you need to get to the urologist, the radiologist, the eye doctor and the Dollar Store, then get the early-bird at the diner, you’ll be in love with your automated car. Even if it drives through the plate glass window at the strip mall, it won’t have done any worse than if you were at the controls.

  19. Old Prude says:
    @The Alarmist

    That will be done by the lawyers.

  20. El Dato says:

    I use lane hold assist pretty much all the time, and for the most part it gives me the kind of ride one got in early F-111’s in terrain following mode. It’s difficult to trust for more than minor corrections.

    That’s pretty cool.

  21. the kind of ride one got in early F-111’s in terrain following mode

    Were you joking here? The F-111 was notorious for crashing when it was tested at what was supposed to be its specialty — low-altitude all-weather flying.

    • Replies: @El Dato
    , @Achmed E. Newman
  22. Why do airplanes have automatic pilot systems anyway? Are there aspects of flying that machines are better at, than humans (highly trained humans at that)?

    I’m serious here. When it comes to cars, there’s nothing they’re actually better at than humans. Ok, maybe parallel parking and holding a constant speed. But by and large, the value proposition in autonomous cars is that they free the driver to take his mind off the road. Pilots, however, aren’t commuters with “better things to do”, pilots are already at work and paid to be engaged in flying the plane.

    Maybe there’s a pilot here who can clear this up. I’m (obviously) not a pilot. But I do know this about automatic pilot systems (learned from Patrick Smith’s _Ask the Pilot_): that the “automated pilot” is not a single thing, but a collection of specialized tools.

  23. El Dato says:
    @International Jew

    Pilots, however, aren’t commuters with “better things to do”, pilots are already at work and paid to be engaged in flying the plane.

    Yeah but in olden times you had six people in the cockpit whereas now there are only two (one of which may be taking a nap), traffic is way denser, flying through the night or storms frequent, planes are faster so events and handovers are faster, the plane is more overall more complex and optimized (i.e. behaves in difficult ways in edge cases) and the effort to get rid of “human error” makes automation an good proposition.

    The bare-knuckle ride of yesteryears? Over.

  24. Jack D says:

    The real killer for turbine powered cars is fuel economy. Turbines burn almost the same amount of fuel at idle as they do at full throttle. In a plane this is not an issue because you are running full throttle most of the time (no stop signs in sky) but in a car it is.

  25. @International Jew

    Holding altitude to within 100 feet for hours at a time on a fast plane can get tedious, IJ. It can leave not much room left for other thinking, navigation, communication, etc. It’s one thing on a 120 mph Cessna, which doesn’t often have an autopilot and shouldn’t need one.

    It’s not like airplanes aren’t stable in general (unlike helicopters), but they will drift off of your desired state within a few seconds to a minute, if you don’t keep making slight corrections. The need to keep on altitude (or vertical rate or airspeed during climb or descent) and heading or course continuously will be more draining, even with trim controls which may exist (ALWAYS for pitch) for all 3 axes, the heavier and faster the plane.

    It’s a totally different reason for the need for an autopilot than it is for a self-driving car. Obstacles, other traffic, and very tight maneuvers are not the reason for autopilots in airplanes. In fact for navigation, since autopilots can follow a course line made from either the old ground stations through GPS and IRS, flying to where they are going is the easy par for them. It makes one wonder why there isn’t more talk about automating Air Traffic Control, especially in the enroute environment, now that aircraft will be required to have ADS-B (end of this year) and can talk position and speed to each other’s electronics. (It’d be to bad two, because they are good people, with good jobs.)

    (Sorry I did that last homonym thing just to make your day, IJ. ;-}

    • Agree: Jim Christian
  26. Mrs Mute’s new car has autopilot for the highway. It is a terrifying clusterfuck. The automated distance control on cruise is ok, that part is tolerable, but the automatic lane-keeping is another matter. For liability reasons I’m sure, it requires you to keep your hands (both) on the steering wheel. Then the steering wheel is electrically driven left and right as the cameras and software attempt to keep the car between the lanes. So you must loosely (not too loosely) keep your hands on the thing upon which your life depends while it jumps left and right for what appears to be no good reason whatsoever. After about five minutes the thing was shut down never to be toyed with again.

    Drive by wire is scary enough as-is. But to have your steering wheel under software control while you’re unable to take your hands away and distract yourself somehow hoping for the best is nightmarish.

    Stan is keeping his ancient rack and pinion mechanical steering and his mechanical throttle linkage four wheel drive until the wheels rust off. Then he’s going to take the bus..

    • Replies: @Jim Christian
  27. @El Dato

    “Cult of the Dead Cow”? Isn’t that called “barbeque”?

    • LOL: Hypnotoad666
    • Replies: @MikeatMikedotMike
  28. Jack D says:
    @International Jew

    There are many reasons for autopilot but from the airlines’ POV the #1 reason is that a plane on autopilot will burn less fuel just like your car will get better mileage on cruise control than it will with you pumping the throttle up and down.

    You are looking at this backward. In a modern aircraft the computer flies the plane all the time, even when it’s not on autopilot. Airbus’s have been “fly by wire” for decades. This means that the joystick is just like a computer joystick – it’s not physically connected to any cables or levers that move the control surfaces, it just an electronic device that plugs into a port in the computer (in new cars, the gas pedal is the same). The joystick just talks to the computer and makes “suggestions” as to how the computer should move the control surfaces. The computer has certain “laws” built into it and if you make a suggestion via the joystick that the computer thinks will violate the “law” (might put the plane into a stall or a spin), it just won’t do it. Military planes like the stealth bomber are no longer aerodynamically stable – not even the best human pilot could hand fly them anymore. If you tried, it would just tumble like a stone without the computer making adjustments to keep it stable at a rate that is entirely beyond human capability.

    The best way to think of big jets is to think of them like nuclear plants. The pilots are the guys who watch the dials while the computer flies the plane. You couldn’t run a nuclear plant where humans manually controlled each pump and valve and control rod – there are too many of them and it would be exhausting and pointless because the computer is much better at that kind of repetitive stuff.

    The real issue (which led to these crashes) is not that the humans screwed up when the computer gave the plane back to them, it’s that the computer (and the sensors that the computer uses to make sense of the world) should have been better and not gotten them in trouble to begin with. Instead of the computer flying the airplane or driving the car 99% of the time, they need to add (and they can, eventually) a few more sigmas to the AI so it covers 99.9999% of the situations. Humans are still the cause of the vast majority of all driving/flying accidents so if you can get humans out of the system safety will go up vastly. The insurance industry is already wondering what they will do once auto accidents go down by 80% or more.

  29. Peterike says:

    Self driving cars backed up by Ethiopian drivers is probably a bad idea.

    • LOL: jim jones
  30. @International Jew

    Why do airplanes have automatic pilot systems anyway? Are there aspects of flying that machines are better at, than humans (highly trained humans at that)?

    They are way better at a whole host of things. Actually they are better at pretty much everything.

    An obvious example–giving you a smooth ride in turbulence. The auto-pilots can make much faster and smaller corrections than pilots so you don’t get pitched and rolled about. Another is not stalling or flying into stuff, something pilots unfortunately do do.

    Self-driving cars are better or will be better at a whole host of things as well
    — speed control
    — staying in lane
    — maintaining distance
    — braking

    They just aren’t quite better enough at everything to be ready for prime time yet.

    • Replies: @Realist
    , @Achmed E. Newman
  31. “What does it gain me if I have to alertly manage my automated car all the time so that I can suddenly take over and go all Captain Sullenberger?”

    Now apply this concern to 1000’s of commercial trucks roaming the interstates.

    • Agree: Hibernian
  32. Jack D says:

    When public transport (steam trains) were invented, people railed that they were dangerous and said that if you don’t want to walk you should take a horse.

    Our existing system kills literally tens of thousands of people every year and yet people are perfectly accepting of it, but if something new (driverless cars) kills 1 homeless person who has jumped out in front of the car, it’s a big deal and we should take all driverless cars off the road.

    • Agree: Trevor H.
    • Replies: @Realist
  33. Jack D says:

    It’s like being the king of manual shifting – it’s going to be an obsolete skill. At this point cars with automatic transmissions get better gas mileage than manuals so there’s no point to it. You just have to find something else to be good at.

    The direction has been toward greater automation from the very beginning. In the very earliest steam engines (which turned very slowly) humans opened and closed the intake and exhaust valves manually for each cycle. Then someone (legend has it that it was a little boy whose job it was to operate the valves inside an English coal mine but it’s probably not true) figured out that they could rig up some ropes and pulleys so that the engine itself would open and close its own valves. Of course now engines turn at thousands of RPM and it’s completely beyond human capability to manually operate the valves (and to the extent that you could do it, the machine does it better anyway). Manually shifting (and soon manually parking, and manually staying in your lane, etc.) is going to join the list of things that machines do better than humans.

    • Replies: @AnotherDad
    , @Anonymous
  34. El Dato says:
    @International Jew

    File under “Kalman filters can’t do everything”

  35. Alfa158 says:

    The model on the hood looks faintly trans-genderish. I guess that really was the car of the future.
    When I was I kid cars like the Firebirds were supposed to be our future along with free atomic power, vacations on the moon or the bottom of the sea, a cure for cancer, and houses made of plastic.
    The future isn’t what it used to be.

    • Replies: @AnotherDad
  36. Jack D says:
    @El Dato

    ran over a pedestrian walking her bicycle over the street:

    It was at night. The pedestrian was a homeless person (possibly in a mind altered state) who suddenly jumped out from the median in the middle of the block (maybe she had been sleeping there?) into the left lane. It’s not clear whether a human would have seen her in time either.

    But, legally you are right – self driving cars raise huge liability issues. There is a whole industry of PI lawyers who are not going to let their cheese be moved without a fight. They are going to see self driving cars as a bonanza. Every time the robocar runs over the homeless person who jumps out in front of it, instead of poor Latisha the driver who has the state minimum $15,000 policy and no other assets, you have deep pocketed GM and Tesla to sue for millions for every accident and you can afford experts to come and say that the software was no good. Unless they deal with this legislatively, self driving cars are going to be priced like small planes – $50,000 for the car and another $250,000 on the sticker to cover 20 years of future lawsuits against the maker for every car accident that happens from now on.

    • Agree: Johann Ricke
  37. TheJester says:

    I don’t use “cruise control” on the Interstates … because it’s not worth it. My 2013 SUV doesn’t have radar, so I would be constantly taking control and giving back control as I approach slower vehicles. Would radar help? I doubt it. I would still have to pay attention to decide to go slow (too slow) behind another vehicle or pass. I just pay attention and drive.

    Decades ago, I told my teenage son to never engage “cruise control” unless on a highway. There were too many decisions to make and too little time to make them inside a town or city. Two damaged cars, a wrecked undercarriage, and a plowed-up-lawn later … and he had learned his lesson.

    I spent the last 17-years as a software development program manager developing software for DoD … and 28-years dealing with the design of DoD command and control systems. I’m playing with an “Alexa-controlled” home in retirement. My experience: there is no such thing as “bug” free software. There are “certified” software systems such as railroad track control and aircraft control systems, but they tend to be so expensive and time-consuming (the extended testing) that railroads, airlines, and software development firms try to avoid them at all costs.

    Evermore complex computer controlled systems mean ever-longer and more costly testing cycles. I mention the F-35 fighter as an example. Sometimes the software integration problems are simply too big in scale to resolve. They are moving problems that cannot be comprehended nor managed. Truly self-driving cars and self-flying airplanes are among those software problems too complex to comprehend nor manage.

    Both electric cars and self-driving cars are Postmodern “tropes” with loose links to reality. They are political and marketing memes designed to get people to pay premiums as “early adopters” and “unpaid testers” in the S-Curve of a typically “bug-ridden” product development lifecycle.

    The most often-heard phrase in software development, “We’ll fix it in the next release …!”

    • Replies: @Anon7
    , @Achmed E. Newman
    , @JMcG
  38. Self-driving cars will never happen.

    You’re talking about completely replacing an entire existing vehicle fleet consisting of hundreds of millions of cars with more expensive, vastly more complicated vehicles and their associated infrastructure of roadside sensors and bandwidth.

    The capital does not exist to do this.

    The capital does not even exist to maintain our existing roadways or to replace our existing vehicle fleet with equivalent models. In 2018, new vehicle sales actually went down YoY. The system is unsustainable and unaffordable. People are taking out 72-month loans, paying $500 a month for a rapidly depreciating asset, and the whole industry is now dependent on inducing the very last subprime borrower to place himself permanently into hock. This cannot and will not go on indefinitely.

    The self-driving auto is an extreme example of the sort of capital misallocation that occurs at the end of a bubble. In a few more years this whole subject will be a ghost town. No one will be thinking about it anymore.

    • Agree: RationalExpressions
    • Replies: @Jack Hanson
    , @Anon7
    , @Jack D
  39. Back when I were a lad we did “unusual attitudes” during instrument flight training. The instructor would put you in a crazy situation and you had to get the plane back to straight and level by reading the instruments and ignoring your instincts.

    Not that hard.

  40. @Jack D

    I hope that most passenger jets do not run full throttle most of the time. I think that they run at cruise, which is much less than full throttle. Full throttle is kind of hard on the engine, so it’s usually reserved for take-offs or go-arounds..

    • Replies: @Jack D
    , @Inquiring Mind
  41. @Jack D

    Unless they deal with this legislatively, self driving cars are going to be priced like small planes – $50,000 for the car and another $250,000 on the sticker to cover 20 years of future lawsuits against the maker for every car accident that happens from now on.

    Could such a law withstand constitutional challenge? Some state laws limiting liability for other products and services have been successfully challenged as unconstitutional.

    • Replies: @Jack Hanson
  42. @Alfa158

    When I was I kid cars like the Firebirds were supposed to be our future along with free atomic power, vacations on the moon or the bottom of the sea, a cure for cancer, and houses made of plastic.
    The future isn’t what it used to be.

    You got the glory of more diversity! instead.

  43. Hibernian says:

    ” In that scenario, it allows him to do “office work” while the car drives itself in slow heavy traffic.”

    Good way to kill a crazy stupid pedestrian if one appears.

    • Replies: @Hoff
  44. @Jack D

    But, legally you are right – self driving cars raise huge liability issues. There is a whole industry of PI lawyers who are not going to let their cheese be moved without a fight.

    Maybe self-driving cars should be engineered only to run over personal injury lawyers.

    • Replies: @Achmed E. Newman
    , @mmack
  45. Realist says:
    @Jack D

    Our existing system kills literally tens of thousands of people every year and yet people are perfectly accepting of it, but if something new (driverless cars) kills 1 homeless person who has jumped out in front of the car, it’s a big deal and we should take all driverless cars off the road.

    More than one person has died from driverless cars. What is the point of driverless cars??? The down side is it allows the government to track your every move. And they are susceptible to hacking by those wishing to harm.

    When public transport (steam trains) were invented, people railed that…

    Nice pun.

    • Replies: @Jack D
  46. Realist says:

    Self-driving cars are better or will be better at a whole host of things as well

    You forgot allowing the government to control how and where you drive. And to know your exact whereabouts at all times.

  47. @International Jew

    No, no joke there. I’ve been in a F-111 (“Raven” electronic-warfare version of the plane) simulator – no I didn’t fly that plane, but a friend let us try it out.* This was 20 y/o technology, BTW. When in the nap-o-the-earth terrain-following mode, the autopilot HAD to be on. It would beep at the crew to let them know when a big high-g pull-up was coming to keep from hitting the rocks!


    * It ended when I crashed into the tanker plane, haha.

  48. Not quite on topic, as not about car automation, but the previous 737 thread has gotten old:

    I was talking to a Boeing 737 pilot today – he went over some of this. He did not know about the optional indication that will warn when this extra down trim movement due to software (even with the autopilot off) is in progress, as he does not fly this MAX variant.

    The Lion Air crash pilots worked to overpower the down-trim, or just manually (well, it’s electric, but with the usual switch) trim up against it, as the MCAS was working against erroneous data*. There were some circuit breakers that the pilots should have memorized to be pulled in this situation, or for any stabilizer trim runaway. However, it’s one thing to remain calm and collected even with an engine fire or something very bad like that, but if you are in the middle of working to stop this thing from diving toward the ground, then it’s not easy to get your brain in the mode of “what exactly is happening, and what was I taught about this?” An indicator to show that this MCAS is in operation could have saved them.


    * from a dysfunctional AOA vane, from what I’ve read, not per this pilot.

    • Replies: @AnotherDad
  49. @Jack D

    Our existing system kills literally tens of thousands of people every year and yet people are perfectly accepting of it, but if something new (driverless cars) kills 1 homeless person who has jumped out in front of the car, it’s a big deal and we should take all driverless cars off the road.

    The direction has been toward greater automation from the very beginning.

    Manually shifting (and soon manually parking, and manually staying in your lane, etc.) is going to join the list of things that machines do better than humans.

    I agree with your thrust here, much more than the thrust of your comment Steve quotes.

    I find it weird all these people who don’t think driverless cars are possible or the future or goodness.

    The whole reason i’m sitting here longing in my chair, electrons wiggling to and fro down wires and photons zipping across the air to bring me iSteve, rather than out slopping the hogs or working the pasture or fixing fencing–while waiting for the ground to be ready for spring ploughing and planting–is because of automation. Figuring out how to do stuff better. More machine, less labor.

    I think there are issuse with human adaptation–leisure and modernity seems to impact women’s behavior very negatively, worse than men’s–but overall i like being able to enjoy life without working like a beast of burden for my daily bread.

    • Agree: Old Prude
  50. El Dato says:
    @El Dato

    Just working through my reading list:

    Under “Metamorphic Testing of Driverless Cars” (In Communications of the ACM, March 2019), paywalled, we find:

    ON MARCH 18, 2018, Elaine Herzberg became the first pedestrian in the world to be killed by an autonomous vehicle after being hit by a self-driving Uber SUV in Tempe, AZ, at about 10 p.m. Video released by the local police department showed the self-driving Volvo XC90 did not appear to see Herzberg, as it did not slow downor alter course, even though she was visible in front of the vehicle prior to impact.

    Subsequently, automotive engineering experts raised questions about Uber’s LiDAR technology. Li-DAR, or “light detection and ranging,” uses pulsed laser light to enable a self-driving car to see its surroundings hundreds of feet away. Velodyne, the supplier of the Uber vehicle’s LiDAR technology, said, “Our LiDAR is capable of clearly imaging Elaine and her bicycle in this situation. However, our LiDAR does not make the decision to put on the brakes or get out of her way” … “We know absolutely nothing about the engineering of their [Uber’s] part … It is a proprietary secret, and all of our customers keep this part to themselves” … and “Our LiDAR can see perfectly well in the dark, as well as it sees in daylight, producing millions of points of information. However, it is up to the rest of the system to interpret and use the data to make decisions. We do not know how the Uber system of de-cision making works. “

    Regardless of investigation outcomes, this Uber fatal accident raised a serious question about the perception capability of self-driving cars: Are there situations where a driverless car’s onboard computer system could incorrectly “interpret and use” the data sent from a sensor (such as a LiDAR sensor), making the car unable to detect a pedestrian or obstacle in the roadway? This question is not specific to Uber cars but is general enough to cover all types of autonomous vehicles, and the answer concerns every human life. Unfortunately, our conclusion is affirmative. Even though we could not access the Uber system, we have managed to test Baidu Apollo, a well-known real-world self-driving software system controlling many autonomous vehicles on the road today ( Using a novel metamorphic testing method, we have detected critical software errors that could cause the Apollo perception module to misinterpret the point cloud data sent from the LiDAR sensor, making some pedestrians and obstacles undetectable. The Apollo system uses Velodyne’s HDL64E LiDAR sensor, exactly the same type of LiDAR involved in the Uber accident. We reported this issue to the Baidu Apollo self-driving car team on March 10, 2018, MST (UTC -7), eight days before the Uber accident. Our bug report was logged as issue #3341 ( We did not receive a response from Baidu until 10:25 P.M., March 19, 2018, MST—24 hours after the Uber accident. In its reply, the Apollo perception team confirmed the error. Before presenting further details of our findings, we first discuss the challenges of testing complex computer systems, with a focus on software testing for autonomous vehicles.

    There will be dead people on the road to “self driving supremacy”.

    • Replies: @El Dato
    , @Jack D
  51. @anon

    Hat’s off to you. I can’t parallel park for squat!!

    And I sure do miss the ol’ 4-speed on the floor.

  52. El Dato says:
    @El Dato

    From a link in said article, testing showing that neural networks trying to deduce target direction from optical data get confused and might drive you into the greens:

    More from the article:

    Figures … show a real-world example revealed by our test, whereby three cars inside the ROI [Lidar Region of Interest] could not be detected after we added 1,000 random points outside the ROI. Figures … show another example, whereby a pedestrian inside the ROI (the Apollo system depicted this pedestrian with the small pink mark in Figure 4c) could not be detected after we added only 10 random points outside the ROI; as shown in Figure 4d, the small pink mark was missing. As mentioned earlier, we reported the bug to the Baidu Apollo self-driving car team on March 10, 2018. On March 19, 2018, the Apollo team confirmed the error by acknowledging “It might happen” and suggested “For cases like that, models can be fine tuned using data augmentation”; data augmentation is a technique that alleviates the problem of lack of training data in machine learning by inflating the training set through transformations of the existing data. Our failure-causing metamorphic test cases (those with the random points) could thus serve this purpose.

    The scope of our study was limited to LiDAR obstacle perception. Apart from LiDAR, an autonomous vehicle may also be equipped with radar. According to the Apollo website (, “Radar could precisely estimate the velocity of moving obstacles, while LiDAR point cloud could give a better description of object shape and position.” Moreover, there can also be cameras, which are particularly useful for detecting visual features (such as the color of traffic lights). Our testing technique can be applied to radar, camera, and other types of sensor data, as well as obstacle-fusion algorithms involving multiple sensors. In future research, we plan to collaborate with industry to develop MT-based testing techniques, combined with existing verification and validation methods, to make driverless vehicles safer.

  53. @Jack D

    The real issue (which led to these crashes) is not that the humans screwed up when the computer gave the plane back to them, it’s that the computer (and the sensors that the computer uses to make sense of the world) should have been better and not gotten them in trouble to begin with. Instead of the computer flying the airplane or driving the car 99% of the time, they need to add (and they can, eventually) a few more sigmas to the AI so it covers 99.9999% of the situations. Humans are still the cause of the vast majority of all driving/flying accidents so if you can get humans out of the system safety will go up vastly. The insurance industry is already wondering what they will do once auto accidents go down by 80% or more.

    This is very good Jack–spot on.

    I was fixing to write a comment in response to your comment Steve posted, because i think it threw everything into the pot and blurred what was going on.

    In the Air France 447 crash the auto-pilot did not kick control back to the pilots because it threw up it’s hands “this turbulence is too complex, i can’t figure it out–take over humans!” Nor did it turn back control with the plane in a “pecarious attitude”. Rather the plane was flying straight and level and the auto-pilot was handling the turbulence very well–much better than the pilots, especially the junior pilot, ever could.

    The auto-pilot cut out simply because the pitot tube (air-speed)sensor iced up and stopped working, so the auto-pilot didn’t have the data to do it’s auto-piloting thing.

    If there was either
    — a backup airspeed sensor (GPS?)
    — or an additional subroutine in the autopilot to muddle through with the airspeed sensor failure (hey, i’m getting low\zero airspeed readings, but my nose is level and i’m not losing altitude … let’s just keep going!) as the pilots are *supposed* to do
    — or an auto-pilot off stall prevention system like the 737Max MCAS
    … the plane would not have crashed.

    The problem was basically sensor failure, and inadequate backup sensors or complexity of auto-pilot software to keep flying automatically. And then backup system, the pilots (actually the junior co-pilot) turned out to be mediocre. The system was insufficiently complex and redundant to sensor failure and thus too whimpy and accomodating to the human taking control (and screwing up).

    With the Lion Air crash, the sensor–again the pitot tube–failure, but MCAS didn’t recognize it as such, could not distinguish it from a real stall. And did not warn the pilots it was operating or the pilots couldn’t figure out how to disable it. The system was insufficiently complex and sensor redundant, but too cocky and aggressive with regard to the humans taking control (and flying normally).

    But the common point is simply sensor failure–having adequate backups–and more capable software that interprets these sensor failures in the light of other data the way good human pilots do. These are not at all impossible tasks. They just haven’t been done well–yet.

    • Replies: @Jack D
  54. Lugash says:
    @Jack D

    It was at night. The pedestrian was a homeless person (possibly in a mind altered state) who suddenly jumped out from the median in the middle of the block (maybe she had been sleeping there?) into the left lane. It’s not clear whether a human would have seen her in time either.

    In my mind it’s 50-50 that a human driver would have been able to stop in time. It’s about 95-5 though that a non-distracted human would have been able to swerve left by about 18 inches and prevent the accident.

    The homeless lady wasn’t sleeping in the median, she was heading back to one of the homeless encampments that have sprung up in the parks to the north.

    It’s ironic that the death occurred on a low(er) traffic road at a low traffic time compared to the highly adverse areas around ASU that Uber was also operating in. Not to excuse the Uber driver, but if you were going to take your eyes off of the road that was the place to do it. Risks that pop up in safe areas can be worse than bad areas.

    • Replies: @Jack D
  55. Jack D says:
    @El Dato

    There are dead people already, ten of thousands of them every year, millions killed cumulatively during the age of the automobile. US traffic deaths BTW hit a low in 2011, down almost by 1/2 since the early 1970s peak as a result of safety improvements. But they have started climbing again – more drivers, more distracted drivers on their cell phone, more “diverse” drivers, etc.

    If you set a standard that self driving cars can never kill anyone, we’ll never have self driving cars and instead almost 40,000 people/ year will continue to be killed in auto accidents in the US alone. What if I told you that self driving cars would still kill 8,000 people per year but, since humans will no longer be permitted to drive, the # killed by human drivers will go from 38,000 per year to zero? Would you be OK with that deal?

    • Replies: @El Dato
  56. Jack D says:

    The homeless lady wasn’t sleeping in the median, she was heading back to one of the homeless encampments that have sprung up in the parks to the north.

    Welcome to America 2018, Mr. Time Traveler. The good news is that we have robot cars that drive themselves, just like it said we would in the World of Tomorrow exhibit at the Worlds Fair. The bad news is that we have homeless encampments in all of our cities, populated by zombie like opiate addicts. They left out the homeless camps in the World of Tomorrow.

    • LOL: reiner Tor
    • Replies: @International Jew
  57. Hoff says:

    Good way to kill a crazy stupid pedestrian if one appears.

    If you review my post, I was referring to stop/go traffic on the freeway. Not a likely play spot for pedestrians.

    • Replies: @Hibernian
  58. Jack D says:

    In the Lion Air crash it was the angle of attack sensor, not the pitot tube (air speed) sensor. The plane shouldn’t have been flying – the sensor had already been reported bad.

    Since the sensors are the “eyes” and “ears” of the computer, it’s important that they be right and that they have redundancy – that’s part of why evolution gave us two of each. You also need software to mediate what happens if the sensors are giving conflicting readings, etc. Three systems is even better – then you can “vote” and whichever reading gets 2 out of 3 wins. The old aircraft maxim was “2 is 1 and 1 is none”. But having 3 of everything is costly and adds weight.

    Part of why things like GPS don’t get integrated into the airplane is that FAA requirements are very onerous. This is good in that you can’t just hang random low quality junk on your airplane like some teenage hot rodder modding his car. Everything that goes in a airplane is supposed to be tested and verified to a high standard. This is good in a way, but it means that the stuff found in airplanes is sometime decades behind (and orders of magnitude more costly than) the electronics in say an iPhone.

    • Replies: @El Dato
    , @Jack D
    , @JMcG
    , @Anonymous
  59. @James Speaks

    I want computer control to park like Elwood

    I love that scene. How many takes do you suppose they needed?

    • Replies: @James Speaks
  60. El Dato says:
    @Jack D

    But GPS _is_ integrated into the airplane.

    Just for the collision avoidance system:

    TCAS IV uses additional information encoded by the target aircraft in the Mode S transponder reply (i.e. target encodes its own position into the transponder signal) to generate a horizontal resolution to an RA. In addition, some reliable source of position (such as Inertial Navigation System or GPS) is needed on the target aircraft in order for it to be encoded.

  61. Anonymous[233] • Disclaimer says:

    Rideshare companies have dropped the ball by focusing on a fully-autonomous cars. They should be focusing on high-end, small autonomous shuttles, which take you on pre-defined routes between major neighborhoods. West LA to Santa Monica. Downtown LA to Silverlake. Etc. How is this different than a bus? For one, rideshares assume access to a credit card, which filters out the homeless and weirdos that make public transportation a shitshow.

    This would allow incumbent rideshares to expand concentrically in cities, while repositioning their human drivers into the suburbs. The smaller players would no longer be able compete on urban trip prices or response times in suburbs, particularly in a good economy where they will struggle to recruit drivers.

    Additionally, by having a patchwork model of some human-drivers and some autonomous shuttles, it creates “returns to scale” from data, allowing efficient multi-leg trips. Human driver picks you up from your house. Drops you off at an autonomous shuttle that is waiting at the next leg. This could drive down trip costs 30-40%.

  62. El Dato says:
    @Jack D

    I don’t want to set that standard, but the self-driving zombie car has to be reliable, dependable, debuggable, analyzable and fixable. There needs to be enough assurance that insurance can be had, legal certainty can be attained and valid case law can be built. It has to live on roads where there is a high percentage of other self-driving cars. These are large obstacles.

    Just today, the SEC is shaking down VW for a few additional hundreds of millions after it got laden with multi-billion payouts for its defeat device shenanigans. Imagine this multiplied by the inscrutable behaviour of neural network software (did you really not cheat with the training data? did your really consider diversity in the training?). Nobody is going to put zombie car on the road under such circumstances.

  63. Jack D says:

    The down side is it allows the government to track your every move. And they are susceptible to hacking by those wishing to harm.

    Every technology has upsides and downsides – the question is whether the benefits outweigh the costs.

    The government ALREADY tracks your every move. You are carrying a tracking device (called a “cell phone”) with you all the time. How many people refuse to carry them for this reason? And they are also constantly tracking your car’s location with license plate scanners, toll collection passes, etc. And current cars can also be hacked. We already have these problems even without self driving cars.

    • Replies: @Realist
  64. Jack D says:
    @Jack D

    but it means that the stuff found in airplanes is sometime decades behind (and orders of magnitude more costly than) the electronics in say an iPhone.

    BTW, we are seeing this in cars already. Car design operates on a much longer cycle than phone design where new models are introduced every year. So when you buy your brand new car, the electronics for the nav system are already several years behind the latest phones. You can download updated software (although, unlike Google Maps which is “free”, car makers often charge hundreds of $ for updates) but the hardware is never going to get any better unless you replace the head unit (and nowadays the screen is highly integrated into the car – you can’t just swap the radio out of a standard enclosure in most cases). If you keep your car for 10 or 15 years (as many around here seem to do) the electronics are going to be woefully outdated – look at what 15 year old cell phones look like.

    The voice recognition in my 2016 Genesis really sucks and the maps are already falling out of date. Luckily, the car does have “android auto” which basically allows your phone to take over the screen. If I update Google Maps or get a new phone, the screen shows whatever is on the new version – it just acts like a touchscreen monitor. If it was an option, I would have skipped the onboard nav completely and just taken a bare screen with a port to plug a phone in (or preferably just talk over bluetooth). The setup I have is kind of kludgy – there’s no really good place to rest the phone and you have this long USB cable coming out of the dash and getting unplugged at crucial times. The car also has Hyundai’s version of OnStar where the car has its own built cell system that you can (for a price) use to call emergency service, etc. – I let it lapse after the free trial. I can call from my own phone. Car makers have been reluctant to properly integrate phones even though everyone has a phone because they don’t want to surrender a profit center to Google or Apple but it’s going to go that way eventually anyway.

  65. @Jack D

    Jack D.: Listen, I respect your opinions on a lot of matters, as you do seem to know a lot about a lot. In this case, aviation is not your forte. It is mine. Besides your story about the pay and treatment for pilots at the regional airline being out of date by 5 years, you are wrong in this:

    In a modern aircraft the computer flies the plane all the time, even when it’s not on autopilot

    No, your description of how the joystick on the Airbus basically controls the flight director and not the actual controls proportionally is correct. That’s one of the major things I don’t like about them. They are not a pilot’s plane, though they’ve got all that extra room where the yoke would have been.

    However, NO planes just fly without being on autopilot, and NO, there is plenty of stuff that the pilots do WITH the autopilot on that has lots to do with the safe operation and assured proper conclusion of the flight (no matter which modern plane it may be).

    Avoidance of local weather, meaning avoidance of thunderstorms, is done in real time, and it takes a lot of experience to know which clouds are the nasty one, where you will get ice, and this is along with the information from on-board and (more so nowadays) ground radar images. Reports of turbulence, which is not always easily forecast, from other aircraft, are used to decide to switch altitudes or change the route appropriately, and the cabin crew must be alerted when unavoidable turbulence coming, or likely to come. (This last is especially important in those long flights of the heavies, in which there are 300 people on-board, eating meals, some of them with real glass and silverware.) It is guesswork that a computer cannot do.

    There is icing to contend with. Though the wings and often tail leading edges are heated, along with engine cowl inlets, there can be icing conditions that one does not want to stay in long. Having the anti-ice on reduces the performance, and so can be a factor in deciding whether one can even get up higher, for whatever reason (a better ride for instance).

    Changes to the routing needed both for air traffic control and for weather are checked for fuel burn, as keeping an eye on the fuel at landing is always important. There is not a big margin – for good weather at the destination it can be legally as low as 45 minutes worth*. It isn’t always that, but then the extra above that can be used up with a long taxi, a lower altitude for traffic, a hold that came out of nowhere (for New York City).

    None of this stuff, Jack, is decided “by the airplane”. However, that does not preclude something that someone our ages would never consider, but maybe those being born today would: Having all this controlled from the ground. Indeed whatever valve or relay gets moved by whatever switches the pilots move or buttons getting pushed for data entry (for route changes, etc.), can just as well be moved via computer by people on the ground. I’m not gettin’ on it though!


    * No, that’s not just the airlines being cheap, though they will do some of that. It takes about 100 lb of fuel burn just to carry 1,000 lb more, as a rule of thumb. The real problem is, the more fuel put on the more people and/or high-paying cargo must be bumped off. It’s always a trade-off if the destination weather is down.

    • Replies: @Jack D
    , @JMcG
    , @Jim Christian
  66. @Intelligent Dasein

    Agree. There is a small group of SV navel gazers who think that when they’re not fucking up driving at slow speeds in a residential neighborhood they’ll be allowed to do it with CDLs towing HAZMAT at 80 mph.

    Having responded to a number of CDL involved crashes in the last two years I think we could do let cutting the number of CDL licenses issued in half. I can’t imagine adding a robot into the mix.

    • Replies: @JMcG
  67. @Johann Ricke

    This is cute.

    The law means whatever a judge wants it to mean when they want it to mean it. The fact you can post this after a judge decided that its totes okay to sue Remington because of Sandy Hook is amazing, to say nothing of the fact you live in an age of Trump v Hawaii.

  68. @Jack D

    … the #1 reason is that a plane on autopilot will burn less fuel …

    That is just wrong. The autopilot has absolutely no effect on fuel burn. A pilot can fly the same profile with it on or off. As I wrote to IJ, though, it’s a workload to have it off an entire long flight.

    • Replies: @Jack D
  69. Realist says:
    @Jack D

    And current cars can also be hacked.

    Not to the degree self driving will be.

  70. Hibernian says:
    @bored identity

    Seems he’s saying there’s more than enough blame to go around, but, yeah, in some quarters placing any blame on Ethiopian pilots, airline, or government regulators will be (loudly) perceived as racist. Plane manufacturers, in addition to paying more attention to safety in both their designs and their field technical service operations, might consider not selling their planes to some third world airlines.

    • Replies: @Jack D
    , @AnotherDad
    , @Anonymous
  71. theMann says:

    So how well are synchronized traffic lights doing? Something that is orders of magnitude less complex than driver-less cars, and barely works well in the easiest of circumstances, should be a good leading indicator of where we are on driver-less cars.

    Oh but wait, there is the added, and to my thinking, impossible to overcome, other issue: having driverless and human operated cars in the same traffic network is unworkable -as if you had half a peanut butter sandwich and half a bowl of cereal smushed up together – not going to integrate no matter what you do.

    Humans have this amazing neural net called the five senses, which allows us to react faster than conscious thought, especially in survival situations. Not to mention muscle memory, the ability to learn from experience, a herding instinct which can actually be useful in heavy traffic, differing reflex speeds which make individual drivers adjust to their own driving style, fast\agressive to slow\careful, and the eyesight and knowledge to look at other drivers and adjust to their adjustments…….

    Driverless cars have their work cut out for them.

    Then there are other issues – will advertisers like having their billboard revenue eviscerated? Will traffic cops like having their probable cause to stop vehicles removed? Would any major US city tolerate the loss of ticket revenue? And how are auto insurers going to react?

  72. Anon7 says:

    You’ve never tried driving a Tesla with Autopilot. Even in its current state, Autopilot doing the driving is far and away easier on long trips. Put the car in a lane you like, set the speed, and let the car drive itself for an hour or two.

    What sort of person drives a Tesla? Recently, someone asked an open forum question to Tesla drivers, what do you do for a living? A quarter of responses were from former USAF pilots. It’s not a trope, it’s a barely legal computer rocket car. Go to YouTube and watch a Tesla Model S P100D beat practically every car you can name 0-60. Better yet, drive a Tesla and enjoy the instantaneous torque. People who think Teslas are golf carts have never driven one.

    Will cars ever have what Tesla loosely calls Full Self Driving (SAE Level 5)? I don’t know. I do think that Tesla will get to Level 3 on the highway (the car does all the driving, but the driver must remain fully alert) within the next 1-2 years. I understand the criticism of the strategy of gathering data from consumer-driven cars, which Tesla does.

    Tesla is an American company that makes cars in America. Tesla terrifies the entire car industry, not to mention the oil industry, which is why you read so much FUD about it. I respect your past background, but you’ve fallen into the propaganda trap against Tesla.

  73. Demolition Man, the most prescient Hollywood movie ever made, addressed the issue of malfunctioning self-driving cars back in 1993:

  74. @Hypnotoad666

    I love that scene. How many takes do you suppose they needed?

    I’m going to guess one, or maybe two. But how many times did they practice before hand? In other words, set up the cameras, set up some cardboard boxes, then run the bootleg turn over and over until you know exactly where the car will land. Then park the cars fore and aft and do it in one take.

    The real question is … How did they get the car out? (In the movie)

  75. Jack D says:
    @Achmed E. Newman

    I realize the pilot still makes many important decisions – this is in his capacity as executive “captain”, not “pilot” or “helmsman” – on a plane these roles are combined. On a ship, these are very different things.

    But as you say, the executive doesn’t necessarily have to be on the plane. The Air Force flies drones in Afghanistan where the pilot is in Missouri. A reasonable first step would be to have the co-pilot stay on the ground. The history of jet liner air crews is that the # has been reduced over time as automation has increased. You could have a pool of say 10 ground based co-pilots for every 50 airplanes in the air. The co-pilots would be available as needed to the same extent as if they were sitting in the right seat (and they would have all the controls duplicated in front of them – a flight sim but with all the buttons connected to the actual aircraft by remote link) but for most of each flight they would not be needed – when the plane is operating normally at cruise, there is little or nothing for the copilot (or for that matter the pilot) to do. During takeoff and landings, emergencies, while running checklists, etc. the remote copilot would be assigned to whatever plane was in need of assistance. Since 80% of the time, there is nothing for him to do, 1 copilot would suffice for say 5 aircraft. Because there would be less concern with medical issues (there are 9 other copilots in the room) you could extend the retirement age and relax the medical requirements – instead of a 200 hr. “co-pilot”, Ethiopian Air could have had a 20,000 hr older copilot sitting in Missouri. Because the copilot is not worried about dying in the next 30 seconds, he could remain calmer and more clear headed and say, “captain, I think we had better run the runaway trim checklist now.”

  76. @AnotherDad

    A.D., in light turbulence the autopilot won’t necessarily give a smoother ride, but it will hold the altitude and attitude of the plane better than a pilot can do. There are modes of the autopilot made specifically to reduce the gains to where the autopilot will NOT react as quickly and severely as an autopilot, and more like a human would in fact.

    That leads to the fact that in severe turbulence, the autopilot will probably kick off anyway, because it could over stress the airplane were it to try to hold altitude (especially) along with course and heading. An experienced pilot will know which parameters to let go of (as in letting the plane get slammed up even a coupla hundred feet while slowly pitching down) to keep from over stressing the plane). A good pilot will let the heading vary all over the place and just try to keep the blue side blue, in serious turbulence. The autopilot would kill everyone if it tried to hold the parameters required by ATC rules.

  77. Jack D says:

    At least by 3rd world airline standards, Ethiopian is considered one of the better ones. Even when Ethiopia was Communist and allied with Russia, they kept flying Western aircraft and the Communist political authorities allowed the airline to operate on a business-like basis. Like many non-American airlines, rather than going out into the market for experienced pilots, they run their own “ab initio” training system. (Even some of the big European airlines like Lufthansa do this, and often the training academies are in the US where flying conditions are better). In Ethiopian’s case, because they are in a more primitive society, they train not only pilots but flight crew, mechanics, etc. – they run the educational system for their entire workforce.

    I guess because they don’t have the same kind of “farm system” of regional airlines, flight schools, FBOs, charters, etc. that allows American pilots to get their 1,500 hours before they can sit in a Boeing, you have 200 hour copilots. This apparently never bit them before this week, at least not that we know of.

    • Replies: @reiner Tor
  78. @TheJester

    Good comment, Jester. I have never used the cruise control, including numerous trips across the whole country. I’ve got a good way to wedge my toes right into place on the pedal. What gets me is OTHER people that do use it. No, I don’t want to sound like a busybody Socialist here.

    It’s just that these people maintain speed up and down hills, while the truckers and I are doing energy management. Those stupid f__ks in the passing land get almost past somebody, then, instead of just speeding up and getting it done, hold that Speed Limit + 8 (whatever they think won’t get them pulled!) and end up next to a car for 2 miles, blocking the fast lane, then falling back, then doing it all over again next uphill.

    It’s hard to drive with these people, and I wish they’d get on the autobahn in Germany and have some serious driving enthusiast ram them off the road.

  79. Anon7 says:
    @Intelligent Dasein

    Economics provides one of the strongest arguments in favor of self-driving cars. Today, having a reliable four seater car like a Corolla or Focus or Civic costs about $25 per day, which is a lot of money. It just sits in your driveway 22 hours per day, depreciating. Engineers estimate that IF it’s possible to do Level 6 (no steering wheel required) autonomy, it would be possible to provide the same level of service to individuals for about $2.50 per day using public or company vehicle fleets. The endpoint of self-driving cars is not privately owned vehicles, but shared cars. No one who lived in a city would want their own car.

    So is it possible? Consider the elevator “car”, which required an operator until the 1960’s. The reason that it was relatively easy to automate is because an elevator car travels along a track, rather than wandering freely on a road. The track is enclosed in a shaft and not shared with other vehicles or pedestrians, entry and exit is constrained, entry and exit points (the floors) are rigidly defined, the speed is constrained, etc.

    Economics forced the development of the automated elevator. In 1945, elevator operators in New York City went on strike. NYC came to a stop. The strike cost New York a hundred million dollars in lost taxes. It prevented one and a half million office workers from getting to work. Building owners demanded a change.

    This is probably the only way we’ll get to Level 6 autonomy, by having rigidly controlled environments in cities with no human-driven vehicle traffic or pedestrians permitted. I’m not going to say that a mixed autonomous car / human piloted car / pedestrian environment is impossible, but I guess I’ll believe it when I see it.

  80. Jack D says:
    @Diversity Heretic

    OK, so the plane is at 80% throttle when it is cruising – it’s still pretty much wide open and running at somewhere near it’s peak efficiency (and that’s why if you figure out the gas mileage of a jet airliner per passenger, it’s pretty good). But cars rarely (even at cruise) run at 80% throttle and often spend time at idle. In piston engines, fuel consumption declines more or less in linear fashion vs rpm but in turbines it doesn’t so they make lousy auto engines. Nowadays you could use the turbine as a battery charger like the Chevy Volt uses a gas engine or maybe you could shut off the fuel at idle and just keep the turbine spinning with an electric motor and you could overcome that issue but that wasn’t feasible in 1963. But OTOH gas was 29 cents/ gallon and jet fuel even cheaper.

  81. @Clifford Brown

    When it comes to moving freight in volume, trucks are relatively inefficient compared with railroads; that’s why you see mile-long freight trains made of of nothing but trailer-on-flatcar (TOFC) loads for UPS, Fed Ex, etc. etc. Steel wheels on steel rails offers low rolling resistance, and much like an airliner at cruising altitude, once the train is at track speed the engineer’s job is mainly one of managing the forward momentum of the train. The joke among the engineers is that they don’t get paid to get the train moving, the tough part of the job is getting it stopped where it’s supposed to.

    • Replies: @obwandiyag
  82. @Jack D

    I think it’s pretty significant that both 737 MAX crashes happened in a third world country, with all the baggage it comes with – slightly sloppier maintenance, slightly worse pilots, etc.

    That said, it’s pretty obvious that Boeing screwed it up big time. Third world airlines operate thousands of aircraft, and both big crashes just happened to happen to two of the few hundred 737 MAX planes in service. So while third world countries have a significantly worse record, the MAX has a vastly worse record than any other plane in production.

    • Replies: @Jack D
    , @JMcG
  83. Hibernian says:

    I don’t see the word freeway or the equivalent in your original post, but I guess it could be inferred from stop/go traffic in LA. In Chicago we have plenty of parking lot conditions on the expressways but also on local streets anywhere near downtown. Anyway, any other than minimally distracted driving is asking for trouble.

  84. Jack D says:
    @reiner Tor

    the MAX has a vastly worse record than any other plane in production.

    This is a big claim based on an n of 2. It’s possible that Boeing was just extraordinarily unlucky in that two black swan events both happened to their new plane in the same year. For all we know now, even if they had done nothing, maybe the next one would not have crashed for another 20 years. One or two data points is not enough to discern a true trend. You can imagine that you are seeing one or fear that you are but you can’t really tell.

    • Replies: @El Dato
    , @reiner Tor
  85. But that’s supposed to be the pilot’s job – it’s like being a fireman where 99% of the time you sit around and eat pizza

    It’s said that a pilot is overpaid every day of his career, except one. On that day, he makes up the difference, and more.

    Jay Leno explains (2:40) how the Citroen DS was designed so, in the case of a frontal crash, the engine would go someplace else rather than into you. Now that’s my kind of automatic pilot!

    • Replies: @Jack D
  86. Jack D says:
    @Achmed E. Newman

    You are thinking of autopilot in the narrow sense of what you would find on an Beechcraft, where George is holding the heading and altitude that you have punched in on cruise. In a modern jetliner, “autopilot” means a whole suite of things that control just about everything (including the throttle) during climb, cruise, descent, approach, and landing. Once the pilot gets the plane off the ground, he can turn it over to the computer and the computer will fly the whole rest of the preprogrammed route all the way to touchdown (this is not to say that the pilot can go to sleep – he still has to watch the weather and lots of other stuff) Perhaps with a lot of effort you could perfectly duplicate the automated profile and burn the same amount of fuel but it’s doubtful.

    • Replies: @Achmed E. Newman
  87. Muse says:
    @Jack D

    A hybrid car with a turbine generator set might be ideal. A relatively small turbine could run nearly 24/7, supplying enough to charge the battery. If it used compressed natural gas, the turbine would burn quite cleanly. The turbine in combination with the battery would work together for quick acceleration. Power could be generated during breaking as most hybrids currently do. When parked, the battery could charge off the grid if electricity prices are low. If stored outside, the car could dump excess turbine and battery capacity into the grid during peak electricity demand. The trick is sizing the turbine properly for If the battery gets too low, the car would just run at a reduced speed and/or no AC.

    You might even be able to design a thermopile to generate electricity from waste exhaust heat.

    Another plus would be extraordinary reliability. Turbines don’t have reciprocating parts so they don’t require as much maintenance.

    You might be able to use a fuel cell in the same application, but I don’t know much about their reliability or power to weight ratio. Turbines can be quite small, yet powerful.

    • Replies: @Jack D
  88. Jack D says:
    @Reg Cæsar

    Citroen’s had engines in them? Who knew? I though there were hamsters under there. French cars had terrific suspensions – you could drive thru a plowed field with a basket of eggs on the seat and not break any (they had a lot of farmers and a lot of bad rural roads in France). The DS had a hydraulic suspension system which is still better (at least when it is working) than what we have today.

    But owing to an oddity in French tax laws, large displacement engines were heavily taxed and French cars were always badly underpowered. The DS was supposed to be a luxury car (DeGaulle used one as his executive car and survived an assassination attempt despite having a tire shot out because the hydraulic suspension allowed the car to drive on 3 wheels – you can also change a tire without a jack) but it had a 120 cu. in. 70 hp 4 cyl. engine that would not have been out of place in a Ford Pinto.

    In modern cars, the engine is always designed to submarine under the passenger compartment in a crash but the was new in the era of the DS.

    • Replies: @Simply Simon
    , @Reg Cæsar
  89. @Jack D

    I know how all this works, Jack, and yes, the mainline jets have autothrottles. You are talking about the FMS (Flight Management System). It is possible to have the plane programmed for the planned route with an arrival procedure set in, leading into an approach down to an autoland. In practice that hardly ever happens for the reasons I already told you. There’s more going on than you know about.

    Of course you could hand fly whatever profile you have planned. It’s just more work and tedious for anything but a 1/2-hour cruise flight. The fuel burn would be the same. No, all the optimal descent stuff is calculated, can be set in by pilots or ahead of time, but not a single flight goes exactly as planned – ATC complete re-routes, getting held down longer than expected for traffic, the plane does not actually perform as the book says, storms, turbulence, icing, etc. I just went over this stuff. You don’t know as much as I do about this, I can guarantee you.

    • Replies: @Jack D
  90. Jack D says:
    @Intelligent Dasein

    Well, there is probably going to be a lot of sunk cost losses in the auto industry as the industry goes from IC engines to driverless electric. And we are going to need a lot few car factories, repair shops, gas stations, parking lots and garages, etc. – capital is going to get freed up (with some lost). Instead of millions of cars that sit parked 90% of the time, you will summon a driverless uber 2x a day – instead of 2 cars per family you’ll have 1 for every 4. Instead of making $500/month car payments (plus insurance, gas, parking, maintenance, etc.) you’ll spend $300/month on uber rides. This is already starting to happen with young folks even with the uber we have now. Kids are no longer car crazy – they don’t want to own a car any more than they want a landline phone. Like you say, cars are terrible investments but at one time (in addition to being invaluable as transportation) they were status symbols , but now they are losing that and for virtue signallers they are the OPPOSITE – the status is in NOT owning a car.

    But I’m betting that self driving is going to happen – the problem is 90% solved. The last 10% is the hardest but we’ll get there – this is not some passing fad. Our best minds are working on it. If it had to wait for us to redo our infrastructure it would never happen because we suck at that, but driverless cars will carry everything that they need on board.

    • Replies: @Johnny Rico
  91. Jack D says:

    Yes, I was thinking the same (see #76). I don’t know of anyone who is working on this but it would make sense. The power to weight ratio on these things is terrific – a 250 hp turboshaft engine weighs 125 lbs.:

    The only reason they didn’t take over auto racing is because they banned them.

    • Replies: @mmack
  92. El Dato says:
    @Jack D

    Correct and EASA closing the airspace to planes in flight as if Mohammed Atta himself had been in the cockpit was a dick move worthy of fatass haemorroid bureaucracy.

    • Replies: @reiner Tor
  93. @AnotherDad

    The cult of the dead cow doesn’t drink the Kool Aide, it drinks the au jus.

  94. Old Prude says:
    @Achmed E. Newman

    I scoff at these energy management fools. 71 up and down. Deal with it, autotonomous clowns! Who wants to waste brains on the gas pedal, when it can be engrossed in whatever music is worming from the speaker to the mind’s eye?

    • Replies: @Achmed E. Newman
  95. mmack says:
    @Jack D

    The story on the 1967-68 STP Turbine cars is that USAC, the sanctioning body for the Indianapolis 500 and Indy Car racing at the time never outright banned turbine engines. They just added the proviso in 1969 that the engines had to be AUTOMOTIVE turbine engines, sorta like the rules that allowed stock block Chevrolet, Ford, and Buick engines to run inthe Indy 500 up to the 1990’s.

    One thing that killed the turbine was the cost at the time. I’ve seen a figure of $125-150K in 1967 dollars for the Pratt and Whitney engine Andy Granatelli used in the #40 car, versus an estimated $30K for a Ford 4 Cam V8 that had won the 1965 and 66 race. Pop that in the old inflation calculator and that’s $950K vs $225K today.

    • Replies: @Anonymous
    , @Anonymous
  96. mmack says:

    What’s the difference between a dead skunk and a dead lawyer in the middle of the road?

    There are skid marks in front of the skunk.

  97. @The Alarmist

    Were not at least two F-111s lost over Laos in TFM?

    • Replies: @The Alarmist
  98. @Jack D

    When I lived in France I loved the way the low slung Citreon DS navigated curves at thigh speed. I often wondered why they were not exported to the US.

  99. Anonymous[427] • Disclaimer says:
    @Jack D

    The DS and the later CX (which had a transverse engine that rotated crankshaft-right-way) used a Citroen engineered powerplant that traced its dimensions and casting patterns to the 1930s. It was a dependable but underpowered engine though the Brits ported and polished it and added SUs and Webers and warmed it up some.

    Citroen disregarded most every standard for hydraulics, but the hydropneumatic system on the DS and SM was to this day the best such system ever built. It worked well but was a pain in the ass to maintain. All the fittings are odd, metric, and tough to obtain, there is no fitting for a mule, and they changed from a vegetable to mineral fluid midway through the production run and used exactly the wrong colors to designate which is which. (Mineral fluid Cits will work fine with 5606 or in warm climates Dexron but the color is red: the Cit mineral system is green and the veg based system red. In the rest of the world mineral based is blue (or clear amber) and mineral is red. (There’s also phosphate ester that’s purple.)

    Ironically the DS originally was designed to use a flat six aircooled engine. As GM sold Rover the 215 Buick/Olds engine, Citroen could have bought the Corvair engine off GM-it even already rotates the wrong way!, though like the SBC and Detroit Diesels you can make it run either way by changing the cam and a gear or two unlike the Cit engines, but their gallic gallbladder got in the way of that. Insted they partnered with Maserati and built the compact but unreliable V6 90 degree DOHC in the SM, which was and is that car’s Achilles heel.

    Citroen could easily have been as big as Mercedes Benz or BMW in the US market, but they did not want to put success or money over French intransigence. I knew the chief buyer for a line of legendary lawn tractor makers. They decided the 2CV flat twin would be perfect for their new larger tractor and tried to buy them from Citroen. The French did not want to sell them engines but instead of just saying no put them through an incredible amount of crap.

    Trying to do business with French or Russians is such a pain in the ass most people just give up. The French do better work than the Russians but neither wants to be bothered very much and are content as how they are. Citroen was happy with the domestic market and didn’t care what others did or thought.

    • Replies: @Jack D
    , @JMcG
    , @Cortes
  100. Anonymous[427] • Disclaimer says:
    @Simply Simon

    They were from ’57 to about ’73 or so. Bumper and emissions regs ended that but there are always a couple on ebay if you want an exotic, eccentric, but not terribly expensive car to play with. Expect to learn hydraulic repair.

  101. Anonymous[427] • Disclaimer says:
    @Jack D

    Purpose designed automotive gas turbines had thermal feedback in the form of regenerators that took exhaust heat and transferred it back to post-compressor, pre-burner can airflow usually via rotating wheels. Ford, Chrysler, GM, Williams research and Fiat all built running engines independently using this model, Ford actually produced and sold a few truck engines built thus.

    Definitive book is
    The Gas Turbine Engine: Design, Development, Applications by Jan P. Norbye (1975-06-01) Hardcover

    My takeaway from the history of all the turbine automotive efforts is, it’s not as tough as people think to make a turbine engine if you keep it simple, and such aircraft turbines as the P&W PT-6 have an insanely good gross profit margin.

    By the time these efforts wrapped up, the turbine engines were never really competitive in build cost with the Detroit iron piston engine but they did get better mileage and driveability out of these things than one would casually think. The fourth gen Chrysler engine, in the Ghia bodied test cars, got fairly poor mileage in the city but since it would run on kerosene, unleaded ‘white gas’ or diesel fuel it was still cheaper to run than a gasoline car of the time. In fact, its major inconvenience was that one was not to run it on leaded gasoline, that would foul the Cercor regenerator wheels just as it does a catalytic converter. On the highway mileage was pretty good.

  102. Anonymous[427] • Disclaimer says:

    A new PT-6A engine retails now in the half million dollar range, the ST6 industrial versions are cheaper if you can get PWC to sell you ones. Rebuilt and used engines are much cheaper, I have a friend who bought a PT-06T Twin Pac with roughly midtime to TBO hours ut of a wrecked helo for within five figures for a boat project.

    A runout -34 might be in a similar range. My A&P school bought a running but unairworthy early dash number training engine for something like ten grand fifteen or so years ago to meet an FAA requirement. It had probably been pieced together out of old military King Air engines which didn’t conform to the type certificate.

    But-here’s the kicker-Granatelli was not using new engines and paid nothing or a nominal fee for them. The were unairworthy preproduction or training engines, mutts put together from different dash number parts, P&W sent him to get him out of their hair basically. And, with those junk unairworthy engines, he’d have won had things went a little differently. The first 1967 car DNFed because there were no U-joints or flex couplings from the 4WD transfer case.

    (A couple of decades later he took his last junk PT-6 and stuck it in a Corvette: some idiot paid a lot of money for it on the basis it was built around a very expensive engine. It wasn’t. People are forever building cars, and more recently more commonly boats, with junk unairworthy aircraft turbines. The engines can be cheap if they are a dash number no one needs any more. They run around and burn a lot of fuel and make a lot of noise for a while, get biored, and it winds up on ebay for the next sucker.)

    They restricted the “annulus area” for turbines to make it necessary to build a purpose designed one to carry on and made them run on methanol anyway, negating one big advantage of them (kerosene or diesel fuel have much more energy per gallon than methanol) and keeping the Offy Mafia back in the game for another decade.

    My dad knew the guy that built a 4WD twin engine Porsche built car that ran one year. It failed for the same reason most everything but the Offys did: not enough money. With the exception of the Novis, no one else had and would spend the money, if you had the money you ran an Offy in the conventional chassis of the year. But it made for interesting fields, unlike the boring sameness of today. I haven’t paid attention to Indy since I was in high school.

  103. Anonymous[427] • Disclaimer says:

    The last 166 cid Turbo Offys were $65k in the crate minus the turbo and plumbing. At the time a PT-6 was probably retailing new for $200K, but the PT-6 would probably run five or six seasons with zero maintenance. The Offys got “freshened” after every race or second race and fully rebuilt after three or four, replacing pistons, valves, cams , cylinder liners, maybe grinding the crank, maybe replacing the rods.

    Per hour, a racing Offy or DFX Cosworth costs way more to run than a JT8D, let alone a PT-6. Even if you count the fuel.

    • Replies: @mmack
  104. @Hibernian

    Plane manufacturers, in addition to paying more attention to safety in both their designs and their field technical service operations, might consider not selling their planes to some third world airlines.

    That–the no selling part–is a bad idea. It’s better much better if the 3rd world airlines have the most recent planes, with the most recent safety enhancements and the least wear and tear. And as Jack pointed out Eithiopean is the best of the African carriers with a pretty good safety record. In part, by having a decent fleet.

    Working closely with airlines on their service operations is something Boeing and Airbus both do, but–lacking any specific knowledge about their efforts–agree they should probably be working even harder on it and working to make operations as simple and as cookbook as possible.

    But yeah, it’s goodness for these folks to have the latest planes. Much safer than flying something retired from a 1st world major after 25 years of flying. Let that stuff go to cargo, odd jobs or be retired.

  105. Jack D says:

    There was a brief window during the “10 speed” craze of the ’70s where French bicycles (Peugeot, Motobecane) were competitive in the US market, before the Japanese really got going. But every damn thing on a French bike is built to unique French (metric) sizes that no one else in the world ever used – if an English part is 1″, theirs is exactly 26mm, which is ALMOST but not quite the same. Everything closely resembles the dimensions on an English bike but are just different enough not to be compatible – you can tell that they went thru and rounded everything off to the nearest mm just to be French and different and ornery. NO WAY were they going to make parts that were the same as the English or use English measurements – this would have been denying the glory of France or something. You can no longer get French parts for these bikes. Sometimes you can get an English part to fit if you sand it a little or wrap it with tape or something.

    • Replies: @Anonymous
    , @IBC
  106. They don’t care about AI cars. That’s the red flag.

    The matador is AI trucks. Truck drivers cost a lot a big big lot of money. That’s what they care about. That’s what it’s all about. I can just see these endless caravans of robot 48 wheelers barreling down 78. The hair on the back of my neck bristles.

    • Replies: @reiner Tor
  107. @Achmed E. Newman

    You are absolutely correct sir. Professional drivers are on your side. Plus anybody with a brain.

  108. @Achmed E. Newman

    An indicator to show that this MCAS is in operation could have saved them.

    This really seems like the key thing to me. If the anti-stall program is on and pushing the nose down, i damn sure want to know that. It should be a big red light on my console. And i also want a display of the sensor data that triggered it so i can render judgement on whether it’s legit or maybe there’s a problem. And, of course, an easy “push the button Max!” way to turn it off.

    Hate to admit it, but on my first solo flight–which was at different airfield and in a different airplane then i’d done most of my prior instruction in–i was climbing out and didn’t seem to climbing appropriately. I adjusted the trim wheel, and still wasn’t climbing up adequately. I adjusted some more and didn’t seem to be climbing at all. I thought–this is bad. And i’m so young, with so much promise! Then i realized maybe i’d turned the wheel the wrong way. Life got much better after that.

    • LOL: JMcG
  109. @Sgt. Joe Friday

    Of course. You are knowledgeable and exactly correct.

    Thus, trains should carry freight, not trucks. It is common sense, as you so clearly point out. Trains are so infinitely superior in so many ways their superiority is inarguable.

    What everyone’s arguments ignore, however, is graft.

    Trains don’t use enough fuel.

    And that is the only reason there are and will be less and less of them instead of more and more of them. The “free market” working the opposite from the way it should. Since trains are the better mousetrap, the world ought to beat a path to them. But instead, away is the path. Why? Because the “free market” is crooked. The oil industry speaks, and everybody listens. Thus, the worst mousetrap is our mousetrap, until some uppity government gets some guts and makes laws that say no to trucks and yes to trains. Fat chance. But the only one.

  110. Anonymous[427] • Disclaimer says:
    @Jack D

    Are Japanese parts not metric also?

    I know HP went to all metric threaded hardware when they offshored some of their test equipment and Fender uses metric hardware on Japanese (Fuji-Gen-Gakki), other Asian manufacture, and Mexican made guitars, again just enough different not to fit.

    I have in my garage a Velosolex engine made to be mounted on the front of a bicycle. I haven’t even tried getting it going yet, but I already know that to get parts I’m going to have to find someone who speaks French to write the company. It usually isn’t they DON’T speak English, they WON’T.

    With the Japanese, who also tend to ignore English language correspondence, it’s more that despite years of “Engrish” schooling, they really don’t. You want something that’s JDM you get on the jet.

    • Replies: @reiner Tor
    , @Jack D
  111. @Old Prude

    So long as people stay out of the passing lane(s) unless they are really passing, it works out fine. For me, there is no thinking wasted on my right foot. It just does what it needs to.

  112. Anonymous[427] • Disclaimer says:

    The Russians famously made “monkey models” of most of their jets and tanks for export with simplified systems, most people thought this was a scam, but probably they realized many of the Soviet Bloc client states were lower in thinking capacity than Russians (which is saying a lot since there are some dumb Russians out there.)

    Simpler aircraft with larger margins of safety might be better for exports to low IQ nations.

    • Replies: @reiner Tor
  113. @Simply Simon

    I often wondered why they were not exported to the US.

    Uhhh, because they were the ugliest cars anyone had every seen? (This was before the Pontiac Aztek, though, mind you.)

    • Replies: @reiner Tor
  114. @Jack D

    DeGaulle used one as his executive car and survived an assassination attempt

    Take a good close look at that crash dummy. It’s no coincidence.

  115. @Simply Simon

    at thigh speed.

    I don’t want to ask…

    I often wondered why they were not exported to the US.

    The swivel headlamps the company was so proud of were in violation of new (1970s) US federal safety standards. So they bade us au revoir. Other Euro makes did the same, eg, Peugeot, Renault, Fiat.

    Renault stayed active in Canada, so they returned to the US within a decade, with LeCar.

    Once you’re out, you’re out, though. The hardest part of (re)launching a brand is getting a dealer network. After a long absence, Fiat was able to return with a bang simply by buying Chrysler’s.

    Or you can piggyback on a partner–Audi with VW, Mazda with Ford, Mitsubishi with Chrysler. The Colt, the only decent “American” subcompact of the 1970s, was actually a Mitsubishi sold as a Dodge or Plymouth.

    • Replies: @Jack D
    , @Simply Simon
  116. Jack D says:
    @Reg Cæsar

    The Dodge Omni/Plymouth Horizon was originally designed as a French Simca/Talbot (Chrysler owned Simca in those days) although it was built in the US and the US models were considerably different internally. The body styles were highly similar (though not similar enough to exchange panels).

    Peugeot lasted in the US market until 1991 so it wasn’t the headlights. Peugeot is supposedly coming back to the US market “real soon”.

    • Replies: @Reg Cæsar
    , @Anonymous
  117. @Jack D

    It’s possible that Boeing was just extraordinarily unlucky in that two black swan events both happened to their new plane in the same year.

    Well, it’d be technically possible if we had twenty accidents with the first twenty planes of the type. But what are the odds?

    These days, hull losses are incredibly rare events. Even with third world airlines. Hull losses of brand new planes are rarer still, third world or not, and those third world airlines which buy brand new planes are usually among the better ones anyway.

    So we have two hull losses with the same model and with exactly the same type of error, but with two different airlines half a world away. What are the chances of the two being unconnected?

    • Replies: @El Dato
  118. @El Dato

    Yes, it made no sense at all. Those planes had to land somewhere, and they were no more dangerous at the original destination than elsewhere. Besides, apparently only takeoff is dangerous (relatively high AOA, low altitude), and not landing or cruising.

  119. @LondonBob

    I agree. To drive in Boston, or even Huston, requires Will and even some Genius in the true sense of the word.

    A machine that can drive in a city, or get around an unforeseen obstacle in the country, or appropriately park in a suburban neighborhood, is probably an example of strong AI, the machine Antichrist who recursively upgrades its own design. The world would be over before the tech could reach a mass audience.

    • Replies: @Jack D
  120. @obwandiyag

    I guess a glitch will look like this Korean driver falling asleep:

  121. @Anonymous

    It usually isn’t they DON’T speak English, they WON’T.

    That might have been true a couple decades ago. I don’t think it’s still true. Even the way they treat English speaking tourists is now way better.

    • Replies: @Jack D
  122. @Anonymous

    They exported the “monkey models” to Warsaw Pact clients, too.

  123. @Achmed E. Newman

    The Citroen DS? No, it’s beautiful. Like most cars from the era.

    • Agree: Simply Simon
  124. Jack D says:
    @reiner Tor

    I agree. The French seem to have mostly gotten over their “we only speak French” thing, I guess because they realize that the battle is lost and English is the world language that foreign tourists from everywhere (not just the UK and US) speak. They are still fighting the good fight in Quebec where the stop signs don’t say STOP like they do in France.

  125. I’ve been in taxis in Montreal, Quebec and noticed those “ARRET” signs there. They were red octagons alright, but from context (the way the driver acted) I was under the impression that meant “slow down a bit right here”. Were those supposed to be stop signs?

  126. JMcG says:

    Thank you, so very often some amorphous line of thought is crystallized by something I read in this comment section.

  127. JMcG says:
    @Jack D

    I don’t know of a way that GPS could be used to determine airspeed.

    • Replies: @Jack D
  128. JMcG says:
    @Achmed E. Newman

    I was talking to a Concorde pilot once who told me he had to taxi to the terminal at JFK on two engines after crossing the Atlantic. Lack of fuel was the reason. That put paid to any notion I had of coming up with the scratch for a Concorde flight.

  129. JMcG says:
    @Jack Hanson

    You’re not kidding. I honestly believe that truck drivers have gotten much, mich worse since the advent of the CDL. It’s a horror show being on a highway with a semi or a tri-axle these days.

    • Replies: @William Badwhite
  130. JMcG says:
    @Simply Simon

    When I was a lad, the very haughty librarian at my local library swelled around town in a Citroen DS. So at least one made its way over here. I suspect they are like Corvettes in France.

  131. @Simply Simon

    Three were lost in-theatre due to “horizontal stabiliser failure,” which would be consistent with the TFM over-controlling the pitch attitude. Another plausible explanation, since most of our training was done in deserts or over open water, was “scope blanking” due to monsoon rains, which essentially blinded the TFR and might have sent the system into LARA (Low Altitude Radar Altimiter) override mode by default, which would fly the aircraft straight and level into any undetected mountains ahead if the crew was not on top of things. I once heard a non-flying officer suggest the pilots lost their nerve and overrode the equipment and subsequently flew into the ground, but that seems to have vanished from the lore.

  132. @Stan d Mute

    It is a terrifying clusterfuck. The automated distance control on cruise is ok, that part is tolerable, but the automatic lane-keeping is another matter.

    You got right to the nut of it, Stan. Lane-keeping, yeah. Cluster-F is right. I have a 2005 Subaru Outback wagon in addition to my motorcycles. It’s still around because I steer around potholes, depressed manholes, like that. I consider the new cars, then demur, but rent when I have to take long drives to DC or wherever. So, a friend stayed with me last fall for a wedding, flew to Logan/Boston, rented a 2019 Subaru sedan that had a lane-keeping feature in it. What a piece of crap. On the way back she throws me the keys. It’s late-night, dark, on the cratered North Shore roads as all New England’s roads tend to be. Out of habit, I would slide left or right to ‘straddle’ potholes. No dice, I’d cross the center of the road, the damn thing would jerk the car right back into the pothole and WHAM!. Nailed every one all the way back to I-95. I know somewhere in the thing there was a way to turn it off. But can I assume there’s a collision radar of some sort that would allow me to take evasive actions for say, a jogger or a car? Or will it thwart my move to avoid collision and jerk me back into the car or pedestrian because I violated the paint stripe/center lane?

    I don’t know and I wouldn’t want to find out. They really should make a simple on/off switch, turns out it was a program on the screen. They want you to use it, damage to the front end/wheels be-damned. Scroo-em, I ain’t buying it. Besides, it was the slowest car I ever drove. Real shit-heap. Sorry, Subaru fans, I know it’s a cult and all. They’ve outsmarted themselves if you ask me, unless they enjoy picking up the proceeds of potholes, leaving the liability all on the owner? People are fools to run around with this system on.

    • Replies: @Anonymous
    , @Jack D
  133. JMcG says:

    I’m getting like a broken record here, but I’m stunned at the breadth of knowledge on this comments board.

  134. @Achmed E. Newman

    Well, Achmed old buddy, you sound well-versed. Comfort me. When my Jet Blue A320 outta Logan, headed to Tampa spins up, rolls out, rotates to a thousand or 1500 feet, then banks off toward the ocean, who is ‘sticking’ that transition? The computer, or the pilot? At the other end, who is on the stick performing the landing, the pilot or the computer? I get everything is fly-by-wire as regards hydraulics and throttles. But who’s got the actual stick in take-off, climb and landing transitions??

    They had Automatic Carrier Landing Systems (ACLS) aboard Nimitz back in the 70s, but I’m not sure how functional it was then. We still had a crash in May, 1981 one fine Navy night that killed a dozen of my shipmates and broke a BUNCH of airplanes, the fires cooked off a couple of weapons and so on. First bad crash in 5 years I was on that thing. It was all a lark until then. They blamed it on the Marine Corps pilot, his distractions, divorce, whatever. And they took note that forensics revealed many of those killed were weed-dirty, setting off THAT whole mess.

    While I find surprisingly little about that era’s ACLS out on the web, if ACLS was even a thing back then, I suppose it was baby steps leading up to all this? Talk to me, Goose, err, Achmed.

  135. mmack says:


    I have two quoted figures for the cost of the PT6 Turbine used in the 1967 STP Paxton Turbine:

    1) Andy Granatelli in his book They Call Me MISTER 500 on page 289 quotes a cost of $30,000 each for the engines. He compares that to $26,000 for a naturally aspirated (non-turbocharged) Ford 4 Cam V8, and $30,000 for a turbocharged Ford 4 Cam V8. The Ford V8 cost aligns with an add I read in a 1968 Indy 500 program my wife got me that asks “How do race car drivers protect their $30,000 racing engines?”

    2) Roger Huntington in his book Design and Development of the Indy Car on page 140 twice quotes a cost of $100,000, once in text, once in the caption to a cutaway picture of a Pratt and Whitney “ST6”.

    So someone has the dollar amounts wrong, or since Granatelli is writing in 1969, and Huntington is writing in 1981, each author is using time adjusted prices. Huntington writes on page 142 that the work Pratt and Whitney did on the four engines used in the Lotus 56 in 1968 to adapt them to the new 15.9 air intake cost over $200,000.

    • Replies: @Anonymous
  136. Jack D says:

    The Japanese bikes were made for export and for the most part used ISO standard parts. Very different attitude by the manufacturers. Within a few years the rising yen killed their ability to export bikes anyway (manufacturing moved first to Taiwan and then to China) but they (Shiman0) came to dominate the components industry so whatever they make IS the standard.

  137. Jack D says:
    @Boswald Bollocksworth

    I disagree. There are a lot of computer problems that were once thought to be impossible that were later solved and in retrospect were completely solvable once the correct approach was understood (and computer processing power increased). For example, they used to think that optical character recognition was only possible with special (or at least known) typefaces and that there’s no way you could ever get a computer to read a text in an arbitrary font. Likewise, it was thought to be impossible to get a computer to recognize speech from unknown individuals, to translate from one language to another, etc. The key in most cases was to stop taking a “human” approach and rely on brute force statistical techniques. And now these are all being done and not on supercomputers but on pocket sized devices (cell phones).

    Computers don’t have to drive perfectly – they just have to drive better than humans. You could set an initial standard where the computer was better than the best human drivers (take say the accident rate per mile of the top 20% of professional truck drivers) as a starting point for turning them loose on the highway. Over time the AI would only get better from there (while human accident rates are actually going UP in recent years). You have to get past the idea that as inhuman devices, the AI will make DIFFERENT mistakes – they will occasionally cause accidents in situations where a human wouldn’t. BUT, they won’t make human type mistakes (and boy do humans make mistakes) so overall they will drive much better.

    • Replies: @El Dato
  138. El Dato says:
    @reiner Tor

    What are the chances of the two being unconnected?

    One can actually compute that….

  139. El Dato says:
    @Jack D

    For example, they used to think that optical character recognition was only possible with special (or at least known) typefaces and that there’s no way you could ever get a computer to read a text in an arbitrary font. Likewise, it was thought to be impossible to get a computer to recognize speech from unknown individuals, to translate from one language to another, etc.

    I don’t think this has ever been true. It was considered hard to do, which is why there was continual work on getting closer to the goal (I saw some hand-coded processing of scanned writing where we valiantly tried to find enough features so that the result would be correct with high probability in spite of scaling, rotation or shear … it didn’t work so well)

  140. Jack D says:

    It can be used to determine ground speed no problem. Even simple car GPS’s can calculate ground speed – the GPS knows where you are now and where you where a second ago and speed is just distance/time. For airspeed you’d need to know the wind speed too, but knowing your ground speed is a lot better than not having any speed information because your pitot tubes are iced up.

  141. Jack D says:
    @Jim Christian

    But who’s got the actual stick in take-off, climb and landing transitions??

    Taxiing and takeoff are two things that are not yet automated, but climb and landing (usually) are – though Achmed will tell you that there are many exceptions. In Asiana 214, the automatic glide slope was out for scheduled service that day – FAA foolishly believed that airline pilots could still hand fly a visual approach on a clear day so it would be ok to temporarily shut off the ILS.

  142. Cortes says:

    Agreed on the colouring of the fluids. A royal pain in the arse.

    I drove Citroen cars for years and loved them mostly, especially my 2CVs – the ultimate in easy to maintain, and terrific in snow- but fell a bit out of love after a mishap with the suspension fluid in a CX.

    (My usual mechanic had a degree in Mechanical Engineering and used to race 2CVs. He claimed that Citroen used to offer a huge reward to anyone who could overturn a 2CV on a circuit at a Paris airfield without the use of obstacles or devices to tilt the vehicle.)

    • Replies: @Cortes
    , @Anonymous
  143. @Reg Cæsar

    “at thigh speed,” Where was my mind when I typed that?

    • Replies: @El Dato
  144. El Dato says:
    @Simply Simon

    Occupied with the latest Instagram photos of Dame Paige VanZant ?

  145. Anonymous[427] • Disclaimer says:
    @Jack D

    Type certification has not slowed technology down very much on airline and corporate aircraft but it has made most low end GA airplanes very backward in most respects. Piston engine aircraft still hav 1940s/1950s ignition, induction, and propeller control systems that were usable in 1955 for this class aircraft, but not any longer.

    Turbine engine development has been 100 percent sustaining and zero percent disruptive, in the terms used in Clayton Christensen’s book, The Innovator’s Dilemma. Turbine engines are more efficient, use less fuel, produce more power for the weight, and have a longer life and TBO (require less maintenance) than ever but their cost has steadily risen, making them unaffordable to users without extremely deep pockets. A turbine engine that would be disruptive would be like a microcomputer as compared to a mainframe: an Apple II might have cost 1/500th the price of a stripped down least cost mainframe but only 1/1000th the useful processing power. But anyone with a decent job could buy an Apple II. For ten thousand dollars one ought to buy a simple turboprop or turbojet engine. It would get crummy SFC and TBO numbers but it would be reliable (few moving parts) and inasmuch as GA owners often fly only 50-100 hours a year or less-sometimes unfortunately much less-last many years. Large industrial turbochargers have compressor and turbine wheels that are about the size needed: you would need some sort of accessory drive to run an oil pressure and scavenge pumps, fuel controller, alternator, maybe a hydraulic pump.

    That market is studiously ignored. Pwoduct wiability is the first line excuse of course, but then again when aircraft companies run themselves in ways that maximize rather than minimize exposure to judgments that appears to be horseshit. (No one sues the judgment proof or judgment averse. Buying more and more PL insurance is making yourself a juicier target. ) The real issue is that existing customers are more or less satisfied with existing products, no matter how obsolete, and they don’t really aggressively pursue new business. In that respect they are like the French, as Ted Nugent says of deer, thinking only of their next meal or sexual encounter and enabling Nuge to get in with his bow and arrow to whack’em and stack’em. Unfortunately new startups are not very aggressive in GA. Whacking and stacking a Piper or Lycoming would be insanely great, I’d think.

    Some of the experimental guys are using tiny turbines designed for model aircraft, but this results in tiny airplanes with multiple tempermental short lived powerplants.

  146. Anonymous[427] • Disclaimer says:
    @Jim Christian

    The old Subarus were great cars. The new ones are over-tech pieces of shit.

  147. @Jack D

    Thanks. Meanwhile, some guy name of Trevor Sumner tweeted out a dozen twatters today that amount to a fairly damning (and credible to my ear) account of the 737 issues. My specialties were liquid O2 and ejection seats, cockpit AC. I remember technical directives, especially stuff related to seats, all of it flash/critical traffic in to us when it was a Grumman of Martin Baker-issued order. I remember the language they would use. This guy sounds plenty technical. If true, of course. Somewhere, it going to come out, Boeing ought to come clean if there are internal messages that they knew. Maybe they can snow us here, but the international press won’t stand for it. Can’t WAIT to see how this comes out.

    • Replies: @Anonymous
  148. Anonymous[427] • Disclaimer says:

    The ST-6 is the industrial variant of the PT-6. It is known that the ’67 Indy engine has the exhaust housing off a Twin Pac PT-6T .

    PWC is going to charge a lot for any custom work needed for a ST6 variant, but they do offer liquid and gaseous fuel variants and engine power outputs direct from the power turbine or with various gearing arrangements. The PT-6 is a very famous well documented engine and service manuals, all sorts of drawings, etc are on line.

    PWC’s web pages show little or no ST6 info, it appears to be on a if-you-are-serious-contact-us basis, and one no longer hears much about this engine, but the basic PT-6 is in some metrics the most successful “light” aircraft engine in history. It is very reliable, simple, and a pleasure to fly unlike the loud and shaky single shaft TPE331 Garrett, its main rival.

    Salient features include a single centrifugal compressor section and separate PT and CT sections, meaning the engine can turn at up to full power and the input shaft can be at zero RPM. An interesting fact is that since torque isn’t very high at zero RPM, a stout man can grab and hold a still propeller blade and so long as he doesn’t let go that airplane isn’t going anywhere. If he slips and lets go with the engine much over idle, he will probably get sucked in and chopped to bits, making this a usually very dumb idea.

  149. This seems to me, wallowing in my ignorance as I am, to be a pretty damn good explanation of what went wrong.

    • Replies: @Jack D
  150. Anonymous[427] • Disclaimer says:

    The 2CV was great on uncongested rural roads and in town but a death trap on fast highways, particularly the Kansas turnpike with its intermittent high crosswinds.

    The CX had a similar hydraulic suspension to the DS but it was never quite as well loved. The CX engines were reliable and the car drove okay, not nearly as quirky as the DS. There was also a GS that never was sold here and is a cult car as well. It had a VW-like flat four aircooled engine, except for a tiny number that had a Wankel rotary.

    • Replies: @Cortes
  151. Anonymous[427] • Disclaimer says:
    @Jim Christian

    You were an AM-E then I take it?

    I would have been an AM or AD but the recruiter jacked up, so I wound up in the Army instead.

    • Replies: @Jim Christian
  152. @Jim Christian

    Jim, as your JetBlue Scarebus* climbed out of Logan, whoever was flying (it’s usually one leg for the Captain, then then one for the F/O, etc.) had the rudder pedals and the joystick under his control from stopped at the beginning of the runway through the takeoff. On the climb, the autopilot could have been turned on as low as 500 ft. above the ground to way on up in the climb into the flight levels (> 18,000 ft) depending on the mood of whoever was flying.

    On your landing down in Florida, the same guy would switch off the autopilot also anywhere from 5,000 ft or even higher, when cleared for a visual approach, or as low as 200 ft above the ground (less than a mile from the runway numbers. In general, in instrument conditions, not many will hand fly it in until breaking out of the clouds, but there’s nothing that says you have to. It’s ALWAYS good to do as much hand-flying as you feel comfortable with, but if stuff crops up that needs attention (even a small maintenance item), normally the autopilot is kept or switched on (unless the maintenance item IS the autopilot itself!).

    While even the biggest hydraulically actuated-control aircraft have interfaces (rudder pedals, and the yoke – push/pull to move the elevator, and turn for ailerons/spoilerons) that allow for proportional movement of the surface per control deflections or forces, the Airbuses are not quite the same. When you move that joystick, of course it’s not moving a cable or anything, but that’s not the difference. The difference is that instead of just moving the controls proportional to your hand-movement, allowing you to just plain fly, your movements are converted to a change in the position of something called the “flight director” (just part of the electronic display nowadays, but formerly a triangle in front of the artificial horizon moved mechanically from signals from the navigation electronics and other stuff). It’s like your input is telling the plane “get into this new attitude” rather than simply “more roll to the right and pitch up a tad”


    * Sorry, that’s just a habit. It’s not like I wouldn’t ride on one.

    • Replies: @Jim Christian
  153. Jack D says:
    @Bill Jones

    The Gates article from Seattle was also very good. Bottom line is that Boeing lied about the criticality of the AOA sensor so that they could get away with having only one and now hundreds of people are dead. Originally MCAS was only supposed to command .5 degrees of elevator pitch but in the end it commanded 2.5 degrees but they never told the FAA.

  154. @Jim Christian

    I don’t know that much about the military aviation world. That time in the F-111 Raven sim. was just when a flying friend let us get on the base to do it, pre 9/11 of course. Your experience on that carrier probably includes a lot of knowledge that I don’t have. That was really something about the crash.

    Back to your question, I was surprised that even the fairly new Boeing 717s, that are actually McDonnell-Douglas planes in reality, built in Long Beach and all, have autoland! Before recently, it’s only been the bigger birds, for reasons, I believe, that I mentioned in another comment. As I said though, auto-land is more work than just flying the damn plane, so it’s not the normal mode.

    BTW, Jim if you know that very basic requirement for Private Pilots on up about currency, 3 T/O’s and 3 landings within the last 90 days to be able to carry passengers you’ll appreciate this: For a > 8 hours-scheduled flight there must be 3 pilots on board so that no one sits in the seat for more than 8 hours total (there’s relief in the rules, of course, for a flight that goes on longer than the schedule says). For over 12 hours-scheduled, there must be 4 for the exact same reason. OK, who gets to do the one landing every 8 or 15 hours then? One of the other pilots does one on the other end. Because the trips are so long, pilots may only do 6 or 8 legs a month, and depending on who wants the landings (based on who’s the boss, basically), a pilot will often get sent back to the simulator to get current.

    Haha, if you are a private pilot, I’m sure you’re laughing at that. Between contractual pay, the simulator time itself, and hotels, etc. it could cost the company many thousands of dollars to get a guy current, while you can spend $15 to go up and do 3 touch-and-goes!


    PS, don’t forget to make them full-stops at night to get night-current.

    • Replies: @Anonymous
    , @Jim Christian
  155. @Jack D

    Here’s the thing about the Asian crash, Jack. I don’t still have all the details in my head, but yes, these guys should have been able to do a nice visual approach into SFO that day with a perfectly good Boeing 777 with or without the glide slope. I can’t say exactly what Captain Tu Low and F/O Lee Fuk were thinking (sorry, I still crack up at that …), but I do know this:

    The Oriental airline companies are known to be very strict on by-the-book flying over what some people erroneously deride as “cowboy” flying, as in more playing it by ear. These guys were probably very knowledgeable about every aspect of the flight control system along with all the other systems, and there’s nothing wrong with that. However, I doubt they went off the script enough, and just took the plane and flew, enough to be as comfortable as they should have.

    These big planes all have auto-throttles. When they got low and slow, whether the auto-throttles were off, yet they didn’t realize it, or they were supposed to be on, but not working, there could have easily been 2 or 3 seconds going by before their brains went “oh, I gotta do this”, regarding pushing the power way up. It’s just the way it is when you are complacent and the automation systems work almost all the time. If they’d been hand-flying the whole way in on that approach, I imagine it’d have gone much better.

    See, now that brings me back to the experience aspect. American pilots at difficult hiring periods, and unless they were ex-military, were getting hired with thousands and thousands of hours of all sorts. That means they’d flown many planes that, if not being total POS’s, had plenty of problems now and then, and could not get over the weather and out of the ice. They were used to things being wrong, is what I’m saying, and living with it. You’ve seen it in the movies “just fly the plane!”

    • Replies: @Jack D
  156. @Jack D

    Peugeot lasted in the US market until 1991 so it wasn’t the headlights.

    There were emissions and other safety requirements enacted in the 1970s, and it was too much bother to retool for such a small (to them) market. Citroën’s proprietary headlights were unique to them, so of course that didn’t affect competitors.

    Sometimes, though, a crafty company could leapfrog ahead of US regulators. From the SAAB Museum site:

    Two unique Saab innovations are presented on the 1972 models: an electrically heated driving seat and self-repairing (up to 8 kph, 5 mph) bumpers.

    The US government had announced that standard about a year earlier, and all the other manufacturers said it was impossible to meet it in the time allotted. It took the Swedes about six months.

  157. Jack D says:
    @Jim Christian

    The Subaru system doesn’t use radar/lidar – it uses a stereo camera mounted above the rearview mirror – if you look thru the windshield you can see them.

    • Replies: @Jim Christian
  158. Jack D says:
    @Achmed E. Newman

    Apparently they misunderstood how the autothrottle worked in the mode that they had selected in the autopilot – they thought it would do one thing (keep the speed up) and it didn’t. And as you said, by the time they realized it, it was a few seconds too late to fix it.

  159. Anonymous[427] • Disclaimer says:
    @Achmed E. Newman

    At one time the Singapore airline bought a Lear of intermediate-for-Lears size (maybe a 45?) so that the crews could use it for landing currency. This has historical precedents: as nutty as it sounds TWA bought, of all things, an Ercoupe for the early 707 pilots to fly and the USAF has used a number of companion trainers for exotic aircraft such as the SR and U-2. U-2 pilots at first flew the Air Force’s version of the straight tail Cessna 310 (called a U-3) and later other planes: SR pilots flew the T-38, which they at least painted black (but shiny black, not the iron ball paint). How useful these “companion trainers” are is a matter of conjecture.

    I’ve long felt that everyone, I mean everyone, wanting to fly anything should start out in a J-3, PA-18, Citabria or the like, a simple tandem tailwheel airplane. The Tiger Moth was also superb for the role, but is much more expensive to build (Cubs were the cost of Plymouths in the late forties!) and unbelievably slow. These weed out the inept and teach the teachable fast and relatively safely-you can ground loop one, but you’ll walk out, and actually getting killed in one takes a special brand of stupidity or ineptitude. They in effect have single lever power control (no prop control, fixed pitch, mixture full rich to run or all off to stop engine) and generally a low compression engine that will eat car gas (as I write this avgas 100LL is, here in Johnson County, KS, three times car gas price: I buy a couple gallons a year now for the Gravely, because it doesn’t go bad over the winter and I like the smell). Fifty hours in that, flying every other or third day, and you now have an instinctive feel for day VFR flight in anything anywhere besides a balloon.

    And being so slow relatively short XCs will eat up a lot of time: you’ll log six hours on a round robin between KC, Springfield and St. Louis. So fifty hours is doable in a short time.

  160. Anonymous[427] • Disclaimer says:
    @Jack D

    The SM-a great car in some ways and a POS in others-had the rotating headlights on Euro models and fixed ones on US ones from day one. I don’t remember if DS ever had the rotating headlights. I do know many SM owners in the US who lived in states where they could get away with it converted their cars to Euro spec. Citroen had a kit which was not officially sold in the US but the dedicated all had a contact in France and would have them dropshipped via Canada.

    But the 5 mph bumpers were a no go-Citroen said it would fundamentally compromise the car’s crash absorbency, which is probably true, and that was the final straw making them officially retire. The DS wound up with a 2300 cc fuel injected engine in the last Euro models.

    Putting any other engine in a Cit was made very difficult by the reverse rotation thing. In retrospect a Chevy II or Mercruiser four-or a SESCO- would have worked, as they are easily set up reverse rotation, but the transaxle was not known for taking a lot of torque.

    The DS had a great top speed but poor acceleration, and most drivers interested in sports cars preferred a floor shift rather than the dainty column shift on the DS. American ones usually did not have the Citromatic, which was a great advantage, it was a pain in the ass. Cit buyers in the US were an odd lot, to say the least, but it was a technically interesting car. When old many of them did wind up in the hands of aircraft mechanics, who were not afraid of the hydraulics (but did find them a little frustrating due to the nonstandard everything. )

    • Replies: @Jack D
  161. @Diversity Heretic

    Jet planes are most economical to fly at high altitudes where the air pressure is much lower.

    Did you ever try and cross the Rockies in a non-turbocharged late 70’s vintage 4-cyl car? You floor it and downshift and are being passed by turbocharged Diesel trucks?

    Did you ever fly commercially out of Denver? OK, not only does the thinner air reduce power, but you need to be travelling faster to get the same lift. Maybe the higher bypass jet engines on newer planes are more powerful, but the takeoff roll seemed to take forever as you take on more and more ground speed, yawing from side to side as the pilot works the foot-pedal steering, where you hear from behind the cockpit doors, “Cap’n, I’m givin’ it all she got! She kenna take an’a more!”

    Flying at high altitude is a form of throttling but without the energy losses of a throttle plate. The jet engines develop much more power in the thicker air at ground level for takeoff, but at altitude, they may be at 80% of their max engine revs, but they are developing much less power and consuming proportionately less fuel.

    That is also why short trips on a jet offer fewer MPGs than long ones. You are spending more time climbing out of the lower thicker air that you never quite get back on descent, especially since the engines are inefficient at low power descending into thicker air, and you don’t have time to climb to the really efficient altitudes.

  162. @Anonymous

    They had three ratings under Aviation Structural Mechanic (AM) label. AMH-Hydraulics, AMS-Structures And AME for environmental, escape. AD was broken into ADJ-Jets, ADR-reciprocating. I was always grateful for the rating I got into. Because of the liquid oxygen aspect of my particular rating, we were required to stay very clean as regards oils, fuels and solvents. Because of the ejection seat aspect, I always felt we received a far higher regard from the officers who flew and the maintenance chiefs who drove the mule teams of each shop. Also, AMEs made rank right on time. Others ratings were crowded, I made 2nd Class Petty Officer, E-5 in three years. I knew guys in structures and hydraulics and ADs that were five years and stuck at E4, sometimes E3 through sheer numbers of a crowded rating. They only promoted so many.

  163. @Achmed E. Newman

    Nah, I like the Scarebus, as you put it. All I know is, the thing is at altitude quickly, my three little nips and a cup of ice are in front of me inside of 15 minutes or so, and I have my newspapers via their wireless throughout. It’s all I ask. Until a right and proper 1st Class returns, TSA Speed and extra legroom will have to do.

  164. @Achmed E. Newman

    Haha, if you are a private pilot, I’m sure you’re laughing at that. Between contractual pay, the simulator time itself, and hotels, etc. it could cost the company many thousands of dollars to get a guy current, while you can spend $15 to go up and do 3 touch-and-goes!

    I’m no pilot. Closest I got to pilot is a couple of right-seat ferry flights in the A6 and some skydiving. I came out of the Navy with a fetish for risk. Something had to replace flight decks. For me it was motorcycles, a little skydiving and broads.

    And yes, the final straw of several for me getting out was that crash. Ugly. I finished up my active reserve time at Andrews AFB Naval Air Reserve Unit and bailed out in 1982. I’d seen just enough EEOC by that time to make me a little nauseous, too. I did my duty, but it turns out I’m a lousy puppet.

  165. @Jack D

    The Subaru system doesn’t use radar/lidar – it uses a stereo camera mounted above the rearview mirror – if you look thru the windshield you can see them.

    So are the cameras ‘seeing’ the paint of the double yellow line right before they jerk the car into the potholes? How do they know I’m not moving over to avoid a bicyclist? Also, there’s no value built into the system to allow a driver to avoid a pothole? Hell of a way to run a railroad.

    Thanks for the info on planes.

    • Replies: @Jack D
  166. eah says:

    I’d like to have an automated system on my car for low-speed parallel parking.

    “LOL” — I guess you can handle ‘high-speed parallel parking’ on your own

    The Best Analysis Of What Really Happened To The Boeing 737 Max From A Pilot & Software Engineer

  167. Cortes says:

    I used to commute Glasgow to Edinburgh (overall trip 62 miles each way) in the 2CV. On the outward leg (West to East) with the usual tailwind, I could coax her 600cc up to around 90 mph. The looks on the faces of some people on the inside (slow) lane as we sailed past would’ve gladdened the heart of an ogre. The return leg, in the teeth of the prevailing wind? If you exceeded 55 mph it was only in the slipstream of trucks.

    Happy days.

    (The idiosyncratic column gearshift – as well as the overall quirkiness- meant that despite ease of entry by passers by, the 2CV escaped the attention of thieves.)

  168. @obwandiyag

    Pretty much the truth that the liberal democracy has been consciously anti-train since at least Woodrow Wilson.

    With all that said, the railroad companies benefited from “crony capitalism” (is there any other kind?) back in the day.

    But the railroad companies usually took care of their own. The Pennsylvania Railroad, for one, was a truly great company. Today’s companies are pirates, not builders.

    • Replies: @Jack D
    , @Anonymous
  169. @Jack D

    There is no reason to believe that 10% will ever be solved. Especially since it hasn’t even been defined.

    It has never been demonstrated that in real-world conditions a situation where some large percentage of vehicles are autonomous and the rest human driven that traffic will not be a complete clusterfuck and threat to public safety.

    Try the intersection of Boylston and Mass Ave. in Boston at rush-hour at dusk as a snowstorm starts and the traffic lights get knocked out because of the storm.

    Now add fire and rescue vehicles trying to get through a literal wall of robot cars that will not move because of indecision.

    Have they figured out a way to get robot cars to obey cops?

    Never going to happen.

    • Replies: @Jack D
  170. Jack D says:

    I don’t remember if DS ever had the rotating headlights.

    Yes, it did, at least the Eurospec ones.

    The SM had respectable if not stunning acceleration for a car of its time. The engine (a Maserati V-6 – the French could never figure out how to make a powerful engine) was still small by American (or even German) standards but it was a light car so it worked out. It was a 90 degree V-6 because it started out as a V-8 – not good. But French cars were so quirky to Americans that they had limited appeal.

    • Replies: @Anonymous
  171. Jack D says:
    @Jim Christian

    I think lane keeping mode is supposed to be used on expressways where there are no bicyclists. I suppose there are no potholes either on Japanese expressways. Yes, following the lines is exactly how it works. It’s hard in a rental car because either the manual is missing or you don’t feel like reading it, but if this was your car you would know where the button was to turn off lane-keeping.

    • Replies: @Jim Christian
  172. @Achmed E. Newman

    Cruise control is the second worst thing to ever happen to driving. The first is cell phones.

    I was on I-95 north bound towards Baltimore at 7am on a recent Saturday. Speed limit is 65. Everybody, including the trucks, is going 70+.

    Except two libtards in a Prius in the left lane who are doing 65. The passenger was pouring coffee from a thermos as I passed them on the right.

  173. Jack D says:
    @Johnny Rico

    There is no reason to believe that 10% will ever be solved.

    I think that there is ample reason to believe that. All sorts of problems that were previously thought to be beyond the range of what is possible (e.g. automatic computer transcription of speech by an unknown speaker) have now been thoroughly solved thanks to continued improvements in both the processing power available and a better understanding of AI techniques. Self driving cars is a difficult problem but one with the potential to transform our society in a way unseen since the coming of the automobile itself.

    Once self driving cars become ubiquitous it can be expected that they will communicate with each other and with emergency vehicles such that they will all execute a coordinated ballet to clear a path for the emergency vehicle in a way that is much better than the human response.

    • Agree: Johann Ricke
    • Replies: @Johnny Rico
  174. endthefed says:

    Eric Peters Auto makes a great comment; if you aren’t driving you’re a passenger. It wouldn’t take much confidence in the autopilot and you’re on your phone. Separately, aircraft are generally way better maintained and fly in a generally pristine environment compared to autos. I don’t expect age and wear on sensors is going to make this situation better.

  175. @Jack D

    Well, it wasn’t terribly intuitive. At night is no time to be looking for a button. It occurs to me, another thing I hated about that ‘feature’ was the beep. Women who text and drive need the ‘feature’. Better in the old days when that sort just ran into the back of parked semi-tractors at 75. We don’t really need their genes in the old pool anyway. We’re surely dumbing ourselves down with all the safety stuff.

    • Replies: @Simply Simon
  176. Jack D says:
    @John Burns, Gettysburg Partisan

    Back in the day, many accused the Pennsy of being piratical also. It cut a deal with Rockefeller to give him preferential rates on oil shipments, which had the effect of helping Rockefeller put all of his competitors out of business. It had the Pennsylvania state legislature in its pocket. And so on. It was considered anything but benign in its time.

  177. @JMcG

    For some reason a lot of them are Uzbeks, Kazakh, etc. Just more glorious diversity I suppose. Three or four years ago I was in an Uber. The driver (an Uzbek) was telling me how he used to be a long-haul truck driver but could make more driving an Uber. I’m not sure I believed him (about the making more money part – I figured he’d lost his CDL or been fired or something).

    He told me how he supplemented his truck driving income by helping other truck drivers falsify their log books, allowing them to work longer days. That was his side business – you give me your log books and X dollars, and I’ll pencil whip you into being legal.

    I thought how wonderful is mass immigration…not only are we importing people to do a job that multiple companies are working day and night to eliminate, but as a bonus they’re willing to commit crimes right out of the gate and help other truckers drive while exhausted. I felt enriched.

  178. Jack D says:

    If you are LUCKY there’s a button (I think there is on the Subaru). Otherwise you might find it buried 4 layers deep in the menu structure of the touch screen. The Tesla 3 doesn’t have ANY buttons (or instruments or idiot lights) – it’s just a bare dashboard with a big touchscreen in the middle.

  179. @Anonymous

    The British firm Rover produced a racing model of a gas turbine car in the early 1950’s and again in the early 1960’s. They actually started tooling a production version, P6, of the. 1960’s cars but the project was killed.

    • Replies: @Anonymous
  180. @Jack D

    I appreciate your response. I’m a little more skeptical regarding the promises of future technology.

    I think it was Nicholas Taleb who pointed out that if the technology was possible and someone could figure out how to implement it, it would already exist.

    Read what David Graeber has to say about flying cars.

    I mean, I hope you are right. But this isn’t the only issue that doesn’t make a lot of sense to me regarding a promised future of ubiquitous robot cars.

    I tend to agree with Intelligent Dasein that it will be a ridiculous misallocation of resources as we say goodbye to our happy motoring lifestyle because, well…the shale oil will eventually run out. Haha

    • Replies: @Jack D
  181. @Jim Christian

    “We’re surely dumbing ourselves down with all the safety stuff.”

    How true. I am convinced the insurance lobby is responsible for all the safety devices that have been added to cars in recent years. Take air bags for instance. Prior to air bags a frontal collision could result in serious injuries or death to the front seat occupants. Total payouts by life and automobile insurance companies had to be much higher than if the occupants were only slightly injured because of the air bags. However the personal injury lawyers must take a hit with fewer injuries to settle for.
    My pet peeve is the Tire Pressure Monitoring System (TPMS) These systems need to updated periodically because the batteries in the wheel rim fail or the tires have to be changed or rotated all requiring maintenance or repair at a garage, always time consuming and expensive.

    • Replies: @Jack D
  182. Jack D says:
    @Simply Simon

    The main problem I’ve had is the aluminum stems failing from corrosion. But most units are available aftermarket (Dorman) for not a lot – given how rarely most people check their tire pressure they are worthwhile. Of course if the dealer installs them with factory parts then the cost goes way up.

  183. Jack D says:
    @Johnny Rico

    I think it was Nicholas Taleb who pointed out that if the technology was possible and someone could figure out how to implement it, it would already exist.

    This is like the guy in the 19th century who wanted to close the patent office because he figured everything had been invented already. All sorts of stuff exists now that didn’t exist in the recent past.

    • Replies: @Johnny Rico
  184. IBC says:
    @Jack D

    Jack, in the US in the early ’70s, there would have been at least three different “English” bicycle component standards in common use: Raleigh, BSA/BSC (which I think was eventually adopted by ISO), and also whatever Schwinn and the other American brands may have been using on their older designs. And while French bikes may have had the most comprehensively unique dimensions, they certainly weren’t the only ones that were different: French, Italian, Austrian, Swiss, British, and American bikes all used different dimensional standards and practices that sometimes overlapped but often didn’t. Some examples are Swiss bottom brackets with French metric threading but with the one cup reversed as in British practice, or Italian headsets with ISO interface dimensions except using a different and incompatible thread angle. And remember that some of the smaller parts on British bikes likely used Whitworth threads. Manufacturers love to “screw” around with little (but important) details like these.

    Quite a while back, I remember you saying something about how American adults didn’t ride bikes back in the ’60s or ’70s, and I responded to you with something like “What about the Bike Boom?” And since it sounds like you would have been a young adult right around that time, it seemed a little odd that you would have completely missed that trend even if you weren’t personally involved in it. So it’s amusing to learn that you once took apart your Peugeot 10-speed and then tried to tape the steerer tube when you couldn’t source an original crown race –or something like that. I would have guessed that you’d be more likely to overhaul a Peugeot pepper mill than a bicycle.

    Your comments are often entertaining and incisive, but this one again reminds me of Gell-Mann Amnesia, and perhaps a high-IQ variant of the Dunning-Kruger Effect where smart people draw conclusions from limited data or experience and assume that they must have the right answer primarily on the basis of their own intellectual self-confidence. Here we’re just talking about bikes, but you can see how that can have larger consequences as well (and especially with national or ethnic characterizations).

    And by the way, some of the older French bicycle tire sizes are now among the most widely available in this country: the 700c (622mm) and now the 650b (584mm) sizes have exploded in popularity while older “English” sizes like 27 x 1-1/4″ (630mm) and 26 x 1-3/8″ (590mm) have become almost “ghettoized” with far fewer options available.

    • Replies: @Jack D
  185. Anonymous[427] • Disclaimer says:
    @John Burns, Gettysburg Partisan

    But the railroad companies usually took care of their own. The Pennsylvania Railroad, for one, was a truly great company. Today’s companies are pirates, not builders.

    Consider the GG-1, what a piece of engineering that was. The most magnificent of electric locomotives.

  186. Anonymous[427] • Disclaimer says:
    @Philip Owen

    Rover built the first turbine car, JET 1, but it was crude. They later built several more sophisticated turbine cars and their overall level of success was comparable to Chrysler in the US: Rover DID build and sell a small aircraft and industrial powerplant, the Wolston, and did provide engines for the BRM/Rover racing program. Rover was part of British Leyland and that put the kibosh on actually selling a proper turbine car, as it would have been expensive.

    Either the later Rover or Chrysler engines would have been perfectly suitable for a limited production car. Mileage was poor but in the US no one cared until the 1973 “Oil Crisis”. And again, turbines of this type are fairly omnivorous, running on unleaded gasoline, “white gas”, kerosene or diesel fuel interchangeably. In the late sixties and early 70s diesel and kero were cheap. They would also have been excellent for military vehicles, boats and military and commercial gensets, markets neither company aggressively pursued.

    (Chrysler’s business decisions were also affected by the company’s relationship with Williams Research, in Walled Lake, Michigan: Sam Williams was George Huebner’s lieutenant at Chrysler before leaving the program. WR and Chrysler maintained at least an unofficial relationship throughout the 60s at least.)

    What really sealed the fate of the turbine at least for limited production was emissions: the turbine, originally thought of as “the clean air powerplant” lagged far behind conventional SI reciprocating gasoline engines in meeting emissions regulations deliberately designed to ruin the passenger car market.

    One mistake Chrysler made was not targeting the luxury end of the market, because it had the Imperial, which would have made the ideal platform for a turbine: it cost about what a (standard production) Rolls Royce cost in 1962, Ferraris were not much more. The fourth gen engine was smallish, 400 lbs with its attached TF727 transmission and less than 150 eshp. More like 300 would have been appropriate for the Imperial. It was the sixth or seventh gen engine that made modern accessories like freon A/C and independent power steering and brakes possible with a tertiary accessory turbine.

    The 50 test cars Chrysler loaned out were absolutely beautiful in terms of detail, being coachbuilt by Ghia, but in no wise a particularly high performance car. I’ve sat in a couple of them, the detail is better than any Ferrari or Rolls I have been in, which is a few. Destroying the fleet created a lot of ill will toward Chrysler management for decades. My dad drove Mopars and when we were at the dealer’s I’d bust the salesman’s balls about it all the time. (It didn’t stop me from making a Dodge Monaco my first hot rod project, with a 360 with 11:1 compression, a Clutchflite TF727 and a huge propane tank in the trunk. I showed it to Ak MIller on a road trip -he was in Pico rivera-and he thought it looked pretty good, I was completely gassed (no pun intended) as he was and is my hot rodding hero. The car got totalled by a truck in a thunderstorm several years later, alas.)

  187. Jack D says:

    Your knowledge of bikes far exceeds mine. You might say I am a “Jack” of all trades and master of none. I have an old Motobecane Mixte in my garage which someone gave to me and which I never ride (I keep it for guests) which I have never overhauled – it gets very little use but if something significant were to break I figure it is curtains for it because I won’t be able to get parts (depending on what it is). The thing is heavy as a tank with plain steel tubing and should have been outfitted as a city bike with upright handlebars, fenders, a basket, etc. but it came setup as a “touring” bike like most 10-speeds of the boom, with drop handlebars and no fenders.

    In some cases (mine) the French bikes imported into America did use some parts that were more commonly available in the US such as 27″ (630mm) tires, American pedal threads, etc. so the French (or the American importers) were not entirely insensitive to such issues. Ironic I guess that the 630mm tires are now the harder ones to get.

  188. Anonymous[427] • Disclaimer says:
    @Jack D

    The engine (a Maserati V-6 – the French could never figure out how to make a powerful engine) was still small by American (or even German) standards but it was a light car so it worked out. It was a 90 degree V-6 because it started out as a V-8 – not good.

    A very few DS’s-the DSM-had this engine too, as did a Maserati, can’t remember the model.

    It was a remarkable design, no more than sixteen inches long if I remember, incredibly compact, but it was also very unreliable and mechanically flawed. It was derived from the V-8 in the Bora, which unlike the V-6 variant was a reliable engine. Because of its short size and its wrong way flywheel rotation putting anything else in there has pretty much been a nonstarter. A couple of shops claim to be able to rebuild and correct the flaws in this engine, but I am very skeptical. I would not fool with the SM, just chalk it up as a great idea that didn’t work.

    Had it not been a POS it would have made a hell of a motorcycle engine as well.

    Citroen would have been far better off to buy a modern engine from GM as Rover did, but their Gallic insouciance got in the way. But on the DS, the engine was not the fundamental problem. Poor acceleration, poor rustproofing, and the sheer alienness it presented to mechanics were key downsides, plus the attitude of the US dealer network. BMW and Mercedes determined they would succeed and sell a lot of cars (relatively speaking) in the US, and they did. Citroen and Peugeot didn’t care, and didn’t.

    • Replies: @Jack D
  189. @International Jew

    Are there aspects of flying that machines are better at, than humans (highly trained humans at that)?


  190. @Jack D

    Like what?

    As Taleb and Graeber also point out, the big breakthroughs came as Black Swans and corporate bureaucracy invents little. This is more the situation with robot cars.

    Phones, sliced bread, electricity, refridgerators, AC, microwaves, cars, TV, the internet, planes, robots and a host of other things have been around for a very long time.

    What’s new?

  191. Jack D says:

    How did lopping off 1 cylinder bank change it from reliable to unreliable? I understand that there is a penalty in roughness because it’s now an unbalanced engine but the issues with the engine were stuff that would have been the same on both – the timing chain, the oil pump, the ignition system, etc.

    • Replies: @Anonymous
  192. Anonymous[427] • Disclaimer says:
    @Jack D

    Actually on further study it is not quite the old and proven Maser V8 DOHC engine-it has alternating 2+1 timing chains and is very compact, though the deck height, bore spacing, etc might be common. It was a very unreliable engine for reasons I am not sure of and was the car’s Achilles heel.

    Citroen used wrong way crank rotation because on the Avant and DS the flywheel faces forward, so they used reverse crank rotation so the car could be hand cranked in the common direction. (!) They kept this on the SM and on the CX they went to conventional rotation but the castings needed changing because the accessory section was not designed to permit reverse rotation. So despite being a direct derivative the engines do not interchange. In other words, they forgot nothing and learned nothing. A well designed engine can be set up to run LH or RH with the same basic castings, just different gearing or camshaft or both. GM’s SBC and BBC, Corvair, and inline Detroit Diesels all qualify. (V series Detroits, if memory serves, use different crankshafts. Odd.)

    Amongst production cars I think the only other wrong way engines were Hondas and Corvairs. A Honda V6 might fit in the SM, but I doubt it, as the Maser V6 was very short. Some circle track cars used reverse rotation mostly with the old midget Offy, but that would not be a feasible engine for a road car even if you can find one.

    The Maserati V8 and inline 6 were quite dependable engines, though like most European engines heavy for their displacement and power output. But of course, expensive to rebuild.

    The marriage between the two companies was fairly short lived and the Maserati with this engine is not a particularly desirable one. It was sort of their answer to the 246 Dino, a car I never thought very high of (I remember seeing one on the showroom floor at Gateway Motors on Manchester Road with visible body rust-brand new!) but which brings a ridiculous price now.

    Needless to say. like Objectivists, the Citroen faithful will defend the SM and its engine to the bitter end.

Current Commenter

Leave a Reply - Comments are moderated by iSteve, at whim.

 Remember My InformationWhy?
 Email Replies to my Comment
Submitted comments become the property of The Unz Review and may be republished elsewhere at the sole discretion of the latter
Subscribe to This Comment Thread via RSS Subscribe to All Steve Sailer Comments via RSS
Are elite university admissions based on meritocracy and diversity as claimed?
The “war hero” candidate buried information about POWs left behind in Vietnam.
The evidence is clear — but often ignored
The sources of America’s immigration problems—and a possible solution