Automation and Disaster

Vanity Fair Cover


November 1, 2014

LAST MONTH, Vanity Fair published a story by William Langewiesche on the crash of Air France flight 447, the Airbus A330 that went down in the Atlantic in 2009. You can read it here. Several readers have asked for my thoughts on the piece.

Langewiesche is a fantastic writer and was kind enough to lend a flattering blurb to the cover of my book, so it’s not easy for me to say this: While his article was gripping and meticulously researched, it was painfully disappointing at the same time. The narrative is eloquent and terrifying — the sort of vintage Langewiesche that places him among the very best journalists in America. And yet, when it came to the closing paragraphs, I was steaming mad.

How flight 447 crashed is equally straightforward and baffling: The pilot at the controls overreacts to a faulty airspeed indication and stalls the airplane. Confused by their instrument readings, neither he nor the pilot sitting next to him recognizes the stall, and no attempt is made to recover from it. As the plane descends, essentially out of control, a third pilot — the captain, who’d been on his rest break — enters the cockpit and, quite understandably at this point, can make no sense of the situation. The plunge continues until impact.

Langewiesche focuses on the roles cockpit automation may have played in the accident. His conclusion is a familiar and, to somebody who flies for a living, irritating trope: pilots have become so reliant on automation that we no longer know how to fly. It was a degradation of basic flying skills, more than anything else, that was responsible for the Air France pilots committing such a basic and unforgivable blunder.

But the crash of flight 447 was brought on by a combination of things. Poor airmanship was only one of them. Pilot inexperience was another factor, and possibly crew fatigue as well. Langewiesche also gives short shrift to certain Airbus design quirks that may have played a significant role. For example, the fact that the control sticks are not inter-linked. When first officer Pierre Bonin, sitting in the right seat, first pulled the jet into a stall, the extreme inputs he was making weren’t apparent to the pilot in the left seat. This contributed much to the ensuing confusion.

AF 447 Wreckage

Avoiding aerodynamic stalls, that’s Flying 101 stuff; almost nothing could be more basic. But one point that nobody ever makes is that even in the proverbial old days pilots would, every now and then, make similarly knuckleheaded mistakes, even in old-fashioned, seat-of-the-pants airplanes. The argument in those days was that cockpits needed to become more automated, to engineer out this tendency. The unintended consequence of that approach, we’re now told, is that pilots can’t fly anymore. So what’s the fix? Yet more automation, or less? Or is that the wrong question entirely?

I’m not arguing that pilots’ hands-on flying skill have probably become degraded over the years. But this is because a newer set of skills is required to master the job. A high level of competence is demanded in both skill sets, but it’s unfair, and wrong, to contend this newer set is somehow less important or less valuable than the other.

Neither is it anything easy to learn or master. The most frustrating take-away from the Vanity Fair story is that unless and until something goes wrong, flying modern planes is essentially effortless and without much challenge. The author’s point about the erosion of hands-on airmanship is a useful conversation. However, his contention that piloting jetliners is somehow easy, and his at-times cartoonish descriptions of what the job actually entails, is where the article falls apart (and pisses me off).

Professionals of all kinds will often describe a particular task as “easy.” What they mean, more correctly, is that it’s often routine; they are used to it. That’s not the same thing as easy. Try to imagine how much work — technologically, logistically, and so on — goes into getting a widebody jetliner with hundreds of people on it from one continent to another? It can be very routine, but nothing about it is easy.

Of professional pilots, Langewieschehe says that “a different crowd is flying now, and although excellent pilots still work the job, on average the knowledge base has become very thin.” He cites “nearly universal agreement of this,” but I’ll tell you it simply isn’t true, and there’s a reason why, at least in this country, the major commercial airlines expect a new-hire pilot to have somewhere around 6,000 hours of flight time before they’ll take his or her resume seriously. Because it takes an extraordinary of knowledge, skill and experience to know what the hell you’re doing in the cockpit of even the most modern and “automatic” jetliner.



I put that word in quotes because what does it even mean? The Vanity Fair report is sub-titled “Should airplanes be flying themselves?” The answer is no, and in any case they’re not flying themselves. The entire story starts off under a false premise, and the author, for all of his in-depth knowledge of aviation, does little to dispel the absurd notion, believed today by millions of people, that pilots are merely along for the ride, and the only real challenge to their skills arises in a malfunction or emergency.

No airplane “flies itself.” The automation is only as good as the pilots who are controlling it. A plane “on autopilot” is still under the direction and guidance of its crew. You need to tell the automation what to do, how to do it, and when and where. The automation isn’t flying the plane. The pilots are flying the plane through the automation.

I found it curious that of all the sources Langewiesche quotes in the story, not one of them is an active pilot. As some of you already are aware, one of my longest standing pet peeves is when reporters rely on aviation academics for articles like this — aviation professors, engineers, researchers, etc. Smart as these people are, they tend to have very little understanding of the day-to-day operational realities of commercial flying. Yet they are the ones the media always seems to turn to.

As Langewiesche has it, the piloting profession doesn’t amount to much. At one point he writes of pilots: “All of them think they are better than they are.” I wonder if he’d make such a rude and cursory blanket statement about doctors or other professionals.

At that, at least, I was able to laugh out loud. The point where I had steam coming from my ears came a few pages later: “In professional flying, a historical shift has occurred,” writes Langewiesche near the end of the piece. “Pilots have been relegated to mundane roles as system managers, expected to monitor the computers and sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare even to failure.”

That is about the most asinine and misleading characterization of an airline pilot’s job that I have ever read in my life.


William Langewiesche responds:

I understand your reaction to the piece, and expected it. I want to point out to you that my main sources were senior and active airline pilots, as well as their equivalent among cockpit designers — but because of this, not a single one of them wanted to be identified in print. The broad characterizations of a declining skill level among average line pilots were a surprise to me, and they were universal. They were not my judgement, but the judgment of others in positions to know, and I was taken aback by them. Not once did I invite such characterizations, but they kept coming at me. Moving on, I did not write that the current generations of airplanes are robots just bringing pilots along for the rides, but that the trend seems to be in that direction. The fact of the matter is that crews have been reduced in size and activity as more and more automation has come in over the decades—though the fundamental mental processes of piloting have changed little, and for the best pilots they have probably expanded because they now have to encompass the buried complexities of automation, as well as the basics of flying. Of course I’m aware of the outstanding safety record of U.S. carriers over the past few years. But I was writing about a French crew, in an aviation culture that is decidedly not American, and characterizing airline flying on a global level, just as the airline manufacturers must do. I pointed out that even the global safety record is outstanding. But airliners are not designed for the best crews, or the most thoughtful guys like you, but for the worst, worldwide. Such design decisions inform the experience of piloting for everyone. I tried to make these distinctions very clear, in order to avoid gratuitous insult, or the unfair branding of individual pilots. Obviously this is sensitive stuff, particularly in a proud profession that in the U.S. has long been in retreat, as you know. Anyway I can see why the piece was irritating to you, and as always I enjoy the energy of your response, even when it is directed against me. Keep up the good work, I’ll try to do the same, and I’m sure that in the future we’ll find plenty of other subjects on which we’ll agree.

I appreciate this response and I don’t disagree with it. But I keep coming back to that one paragraph, that preposterous, yet so sure-sounding caricature of a pilot’s job: “…sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare event of failure.” Maybe I’m fixating too strongly on such a small passage, but I have grown so tired of these descriptions. The media repeat them constantly. The notion that flying a jetliner entails little more than sitting there and watching the plane “fly itself” has become fixed in the public consciousness. Millions of people actually believe this. And to hear somebody like William Langewiesche, as knowledgeable and experienced in the field of aviation as he is, reinforcing this mythology through such flip and inaccurate language is exceptionally frustrating. I’m not opposing his main contention, about the degradation of old-school, hands-on flying skills. What angers me is his characterization of the job: the suggestion that it’s easy and involves little more than pushing a few buttons. I guess what I’m trying to say is that flying planes is still pretty complicated and challenging even for the worst pilots.


AirlinePilotHiring Banner Ad

Back to the Ask the Pilot Home Page Visit the Blog Archive Back to Top!

Leave a Comment

Maximum 1500 characters. Watch your spelling and grammar. Poorly written posts will be deleted!

107 Responses to “Automation and Disaster”
You are viewing newest comments first. Click to reverse order
  1. Bilaval Riaz says:

    The fact that captain (the team leader) was not able to sort out the problem would have confused the junior pilots to the extreme
    also at the mid night over the ocean and in the midst of a voilent thunderstorm it would have been a horrifying experience for those pilots to understand the magnitude of situation when thier plane was falling at 10,000 ft per min.and you have about 2 minutes to figure out a problem it was not just enough.
    We should not just blame those pilots but also the design of the airbus A330.and more importantly the C.R.M.

  2. TNT says:

    Hi, excuse my ignorance but if you’re on a plane that’s free falling would you not feel it?
    And why didn’t Bonin know to push the nose down to get out if a stall….? Surely this is 101 stuff? And even if he didn’t know this how did he ever think that pushing the nose up would help? It just seems so hard to believe.

  3. Martin says:

    Looks like the consensus is it was FO Bonin fault as he kept pulling the nose up during high speed stall conditions which is against the book as should have done the opposite.
    The part I have hard time finding explanation for is fly by wire safety feature that prevents pilots executing dangerous maneuvers. you can see videos on youtube where Bruce Dickinson tried to bank Airbus plane too hard or stall it by raising the nose at insufficient speed. Fly by wire stops banking and throttles up to correct pilot errors.
    Don’t you think Bonin assumed that the fly by wire will correct the throttle and have the plane climb by just pulling the nose up?
    I also haven’t seen any resemblance mentioned to the Aeroflot crash when the kid in the cockpit managed to disconnect one part of the autopilot while keep other functions in operations.
    Looks like insufficient understanding fly by wire operation under partial malfunction condition could have been the case of the crash

  4. R says:

    My take-away is that the Airbus interface has more responsibility than is usually considered. Patrick’s one point above, about how the pilot in command couldn’t feel that the second pilot was pushing the stick in the opposite direction, is one important part. Another part is what the Airbus software did when confronted with two opposite and extreme inputs? It averaged them! As a systems engineer, this seems like a decision made by an intern programmer, not a well-thought out response. If the inputs are only slightly different, then either averaging or taking one and ignoring the other may make sense. But when they are diametrically opposed, something else is needed. An alert? An enunciator? Chose the pilot in command’s input? Perhaps all of these.

    • Bob says:

      The basic design of having TWO control sticks that can be operated independently is totally STUPID! There is no way to get around this design error. No matter what you do, there will be a situation that can’t be accounted for.

      1) Accept the pilot’s inputs and ignore the first officer’s? Well, what if it’s the pilot who’s screwing up.

      2) Average the inputs and put up an audible warning? Great, unless you happen to be in a panic situation like happened on this flight. It seems obvious that the pilots never heard the DUAL CONTROL warning because they both kept applying their own inputs.

      3) Put an “override” button on the control stick to lock out the other guy’s control inputs? This feature was present, but since both pilots were pressing their override buttons at the same time, the buttons may as well not have been there.

      In short, there is no good solution to this dumba**ed design. Airbus should never have been allowed to put it into production. It contributed to the loss of flight 447, as well as a crash a few years ago of a southeast Asian airliner that was forced by ATC to fly into a thunderstorm. Airbus said that the crew needed better training. What they needed was a better cockpit design.

  5. David says:

    Anyone who holds an airman certificate knows that an ATP is an extremely talented pilot. The fact that air travel is so “boring” today is due, in large part, to the quality and professional skills of line pilots the world over.

    I love modern aircraft, engineering at a level that is almost perfect, but I fly in back with great confidence because I know, the crew is talented, tested and certified like almost no other profession.

    Thanks Patrick, for calling out those who simply do not understand.

  6. Ozarklead says:

    Looks like 8501 is a carbon copy of 447. I imagine the 8501 c/p crew read the 447 article and “no way” they were getting in tops and having the “probes” freeze up!! I understand the “only” direct control from the c/p to any flight control are rudder cables, and that is only to the hyd. cylinders!! Now there is “aircraft design” along the “KISS PRINCIPLE”!!

  7. PrasadK says:

    Hi Patrick, this time though, I think you “protest too much”. I read through the original Vanity Fair article, and i do agree there are some “eye-catcher” phrases and sentences in there; but overall (and i think even you have agreed) the basic premise is that increasing automation has degraded the basic hands-on skills of more recent pilots, to the extent that when it comes to taking manual control, they find themselves at a loss. The scary thing is that such loss of basic skills begets more automation which makes their “de-skilling” even more pronounced. Parallels are found in all industries, for e.g. a car mechanic of yesteryears know how to handle carburettors and fuel injection systems but today’s mechanic simply plugs an electronic digital device into the car’s ECU and issues a set of commands. If something does not work, he yanks it out and puts in a new one. The car’s designers have completely “de-skilled” the mechanic.

    • Bob says:


      Your comment about mechanics doing diagnostics by plugging in a code scanner and then replacing a defective part is quite accurate. However, you have to give the mechanic a break here a little bit because of the complexity of today’s cars.

      A few years ago, my wife and I bought a 2004 Cadillac DeVille, used. I was fortunate enough to find a Factory Service Manual on DVD on eBay. I’ve since had to troubleshoot a couple of problems in the car and after reading about the various systems and how they work, I can tell you that it often would be impossible for the average mechanic (even an “old time mechanic”) to just take out the books and start trying to figure out what was wrong. And I’m an electrical engineer and understand all of the electronics jargon, data buses, etc.

      For the last five years that I worked for Lockheed Martin, I was responsible for redesigning some of the avionics on the A-10A Warthog and I would have to say that our 2004 Cadillac DeVille is more complex than the A-10A that we were upgrading.

      Of course, now that the A-10C has all of these new electronic upgrades it will be more difficult for the service crews to diagnose and repair problems than it was before. It seems that increased functionality and complexity go together. Knowing where to stop is the key.

  8. User Hostile says:

    Kudos to both you and William Langewiesche for not engaging in snark and remaining polite to one another. Given the polarization everywhere else, it’s quite refreshing to see an attempt to by both sides to honestly understand each others position.

    Really, it was that refreshing.

  9. Bill Palmer says:

    The characterization of an airline pilot’s job as “…sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare event of failure.” is akin to saying an authors job is to type.

    One must not get lost on the physical motions needed to carry out the job. Knowing WHAT to type and when, and when to “intervene” is the part that requires the knowledge, understanding, and experience. As far as “intervening,” a pilot would only be intervening in what he had told the autoflight system to do in the first place.

    Knowing when to intervene, requires all the knowledge needed to fly without the automation.

    • Patrick says:

      >> The characterization of an airline pilot’s job as “…sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare event of failure.” is akin to saying an authors job is to type. << I love that. it's creative... and it's accurate! Thanks for posting.

  10. Jeffrey F. Burr says:

    I liked the response from William Langewiesche, but even for someone who has obviously thought it through and is aware of the pitfalls of overdoing it, it’s amazing how unaware he and other writers are about how the reader is going to interpret glib statements like the one, quoted above, about how pilots do almost nothing these days. That is, they’re going to take them literally. Langewiesche might realize that it’s still a complex job, but that paragraph gives a very different impression to the average reader.

    People commonly voice the idea that airliners can “fly themselves,” and in many cases it’s clear that they mean this literally, or nearly so. I’m not saying that each of these people got this notion from the story in Vanity Fair — but it doesn’t help.

  11. Marcio V. Pinheiro says:

    As a passenger, every time I enter in the plane I try to see how the Captain looks like. From my perspective I am at his hands. Period.

  12. Simon IOM says:

    I wonder with regard to the automatic ” anti stall” computer systems on the Airbus, If they were not present, would the aircraft when it entered the stall have suddenly banked hard over and dropped alerting the crew of a definite stall, would the aircraft maybe have entered a spin though and been uncontrollable anyway or is there a chance of recovering such a scenario from that altitude?

  13. mjg says:

    I have not read the article, but the comments indicate that many readers are discussing the AF447 pilots’ inability to recover from a stall. The conventional recovery is to pitch down and apply power to accelerate the airplane to a speed that provides enough lift to resume normal flight. Indications are that the PF kept applying power but could not accelerate to regain aerodynamic control.

    That indicates a stable deep stall, a nearly unrecoverable condition. The airspeed is so low, [or even zero] that there is no airflow over any control surfaces rendering them completely useless. The airplane can be wings level [roll angle zero] and at level attitude [pitch angle zero] but the flight path can be nearly straight down [very high angle of attack and rate of descent] without hardly any forward speed. The airplane may be in falling leaf motion, but if it does not pitch or roll out, it is doomed. Impact will be hard, flat and non-survivable.

    How did AF447 get into a deep stall? That’s one for Airbus’s aerodynamicists to answer. The phenomenon was first noticed in the early 1960’s on the first rear engine t-tail jetliners like the BAC-111 and the Trident. The A330 has wing engines and a conventional tail, plus lots of automation to stay out of this flight regime. However, blocked pitots can under some conditions bypass or disable the safeguards, which is why pilots must always FLY THE AIRPLANE not manage the flight computers.

    Any thoughts on this from Airbus and/or Boeing engineers and pilots?

    • Well, what you’re looking at is an evolution like this (first observed in military aircraft):

      You start out with an objective: in the case of air superiority aircraft it’s maneuverability, in the case of Airbus it is fuel efficiency. The aerodynamicists tweak elements of the configuration to achieve it, and succeed. But oops, we’ve just shrunk the flight envelope to the point where we can’t trust a human to keep the airplane in it: one sneeze and you’re in a stall or spin, and coffin corner is measured in fractions of a knot (I exaggerate, but you get the picture). Aha! Let’s use a computer to keep the aircraft inside the envelope. And it works.

      And the next generation gets a more capable computer which permits smaller envelope and more performance, and so on. Until you reach the point where the flight control system is there to defend the aircraft from the pilot, whose inputs become “suggestions” instead of “commands.” Really cool, until you get shot up or a pitot tube freezes and the computer goes nuts.

    • Bill Palmer says:

      I’d like to address several aspects of your comment.
      First, I believe that the pilots DID react to the stall warning – which sounded for 52 seconds before the airplane started its deeply stalled descent.
      They applied full power-just as they had been trained to do when they checked out in the A320 years before, which was at low altitude (approx 5000′) (yes, they got no stall training in the A330- and none at high altitude). At 5000′ this is highly effective, at 38,000 feet it has no effect at all.
      The “deep stall” that is referred to does not apply here. That term is mostly a T-tail phenomenon (as mentioned). The AF447 aircraft developed very high angles of attack (according to Airbus engineers I talked to – up to 60°)- (don’t confuse angle of attack with pitch attitude). It was maintained there by a stabilizer trim that had gone to the full nose up position. It probably was recoverable, but would have taken many thousands of feet to do so-probably at up to 10,000 feet based on my own simulator trials.
      When forward stick was applied by the pilots, the airplane DID react. Unfortunately, it was not held long enough for the recovery to take place. Reduction of the full nose-up trim would also probably have been required to complete the recovery.
      The descending flight path was much as you describe (and I too in my book “Understanding Air France 447”) a falling leaf. A downward flight path angle of 45°, wings rocking, yet pitch only at about 10-15°. Indeed with the airplane banked to the right, and full left aileron control input (the roll control was a direct relationship at that time) it only acted to further stall the right wing.

      The situation was partially brought about by the neutral stability of the flight control law the airplane was in when the 3 pitots were clogged: Alternate law 2- without protections. This resulted in the loss of a nose-down pitching tendency at low speed/high angle of attack. Thus, when the airplane was pitched up, it had no tendency to want to pitch back down on its own to prevent a stall. This is usually not a problem as there are several layers of protections that keep the airplane within the normal envelope. In this case, the pilots needed to push forward on the sidestick for as long as it took to pitch the nose down and reduce the angle of attack. The longer the event went on from the initial seconds the longer the input would have needed to be made. At some point, there was not enough altitude to complete the recovery. As the passed 10,000 feet and the captain had finally figured out what was happening and instructed “no, no, no don’t climb” it may well have already been far too late to recover.

  14. Otavio Lima says:

    I think that there is a hidden premise that makes the situation much worse, and that is the belief that technology can substitute for inadequate piloting skills. I think that this is the premise that leads Langewiesche to make flip remarks about flying airliners by using keyboards. I think you came very close to identifying this hidden fantasy in describing the actions of the co-pilot which cause the airliner to stall. You refer to stall recovery as flying 101. If so, why is it that three professional pilots failed to identify that the airplane had stalled?

    • Bill Palmer says:

      >> If so, why is it that three professional pilots failed to identify that the airplane had stalled?<<

      I believe that the two flying pilots did recognize the stall—after all the stall warning sounded for 52 seconds straight. When it sounded initially, and then later, they applied the recovery procedure they remembered: FULL POWER. This was quite effective at the low altitude they were trained at. At 38,000 feet (where the stall occurred) it has vitrually no effect. This is something I often demonstrate to the FOs I fly with—by applying full power at cruise and see…no changes (at cruise, we're at just about full power anyway). I believe this lack of response may have lead to the crew's bewilderment when they said "But, we've got the engines, what's happening?"
      By the time the captain arrived back in the cockpit, the stall warning had silenced (due to the low indicated airspeed), the airplane was descending through the altitude it had when he left, the nose was pitched up, the power at full, and the airplane appeared to be losing altitude (indeed it was). This almost certainly presented a confusing picture which at first doesn't seem possible that all of those indications could be true at the same time. The FOs flying the airplane told him "we don't know what's happening- we've lost control of the airplane." Both true statements.
      For a more in-depth exploration of all of these factors (and others) please see my book "Understanding Air France 447."

  15. Jim Mc says:

    Thanks for a very interesting discussion. I, like others who have commented here, appreciate the respectful way most everyone treats comments they don’t necessarily agree with. I have lost patience with discussions like this that quickly become snark-fests.

    The fact that jumps out to me, after reading both the VF article and Patrick’s comments, is that at the time of the accident Mr. Bonin had less than half the hours that Patrick says are needed to even be interviewed for a job with an American carrier, and very few of those hours were as pilot- in-command. There is no substitute for experience and its hard to see how the ab-initio training mentioned in several of the comments is as good a foundation as the up from the bottom training, grueling as it is, that Americans have to endure.

  16. Bill says:

    Hello All,

    I found the article in VF and the many comments I have read very interesting. I believe there is validity in both.

    For me, the best thing about the article is it has created a discussion both pro and con concerning the items stated in the VF piece. Discussion is good, it creates awareness and by stating thoughts, facts, and opinions we have to articulate both verbally and in writing our thoughts.

    This discussion, when done both emotionally and logically helps to affirm or question our beliefs. Reply comments (discussion) may pose both pro and con statements.

    The civility within the comments on this page is very nice. Respectful and an implied agreement to disagree at times, seemingly without ad hominen attacks. Very rare on any blog and very impressive..

    This says much about the commentators.

    In all, I believe the author has done anyone associated with Aviation a service. By association I refer to pilots, aeronautical engineers, passengers and the curious, etc…

    Thank you

  17. MS42 says:

    We are nearing the age of autonomous cars (Google Self-Driving Car). I’m guessing that when it becomes commercially available, somebody in Marketing will call it ‘Chevrolet Impala AutoPilot’. 🙂
    It’s natural that the driving public will equate the two.

  18. Air France Captain says:

    THANK YOU for the comments on the Vanity Fair piece. Your analysis helped me control some of the steam that had built up in me over the course of that story.

    I’m an Air France captain, and I was a close friend of David Robert, one of the two first officers of AF447.

    There is something that was never brought up by any media, but that, in my opinion, played a pretty big part in the outcome:

    I don’t know if you’ve ever flown an Airbus from the A320/A330 family, but I remember very well the first thing that was said to me about this aircraft when I started my A320 type-rating back in 1997:

    “This Aircraft DOES NOT STALL, period. There isn’t even a stall recovery procedure because: THIS AIRCRAFT DOES NOT STALL!”

    This notion was engrained in us very strongly by instructors adamant about “selling” the Airbus to pilots whom they knew had many reservations about it. So, to me, it’s not surprising that my colleagues never recognized the stall they were in because… THIS AIRCRAFT DOES NOT STALL.

    Of course we now know better, and we do have such a procedure in our training manuals now.

    • Bill Palmer says:

      The moment any instructor told you “This Aircraft DOES NOT STALL” you should have stood up and called him a liar! As a pilot, any pilot, you should have instantly recognized the absurdity of that statement. It can only be said to be accurate couched within a certain set of conditions – Normal law active, no extreme upset conditions (windshear, etc), related systems working, etc.

      Of course the airplane can stall, as can anything with wings.

  19. --G. says:

    Patrick wrote:

    >>But the crash of flight 447 was brought on by a combination of things. Poor airmanship was only one of them. Pilot inexperience was another factor, and possibly crew fatigue as well. Langewiesche also gives short shrift to certain Airbus design quirks that may have played a significant role. For example, the fact that the control sticks are not inter-linked. When first officer Pierre Bonin, sitting in the right seat, first pulled the jet into a stall, the extreme inputs he was making weren’t apparent to the pilot in the left seat. This contributed much to the ensuing confusion.<<

    Is there a reason why the control sticks aren't interlinked? Would an interlink create a different set of potential problems?

    • Eirik says:

      Interesting question, which I hope someone can explain.

      My guess would be – or at least maybe its one of the reasons;
      Lets say you have a pilot in total panic mode and he refuse to let go of the controls, the other pilot can take command over the controls by pushing a button when they are NOT interlinked.
      Of course, the panic pilot can take the control back, but at least it gives you a chance to cut him off.

      What I dont really understand is how/why both pilots are able to use a non-interlinked system at the same time. They would have to be very coordinated when using the “stick” at the same time, since different inputs will cause the system to cancel each other out. Im not sure how sensitive the system is and how much it takes to cancel each others inputs, but Im guessing it takes less than one pilot going left, while the other decide he wanna go right.

      • Randy says:

        My understanding is that the Airbus will average the inputs. I don’t know why the designers chose to do that. To me, it might make sense for minor differences, but not for major ones.

  20. Joe Cruickshank says:

    I am not a pilot and I do not mean to degrade what commercial pilots do, but after watching USAF pilots “fly” drones from the comfort of their chairs at a base in the Nevada desert, it should be no surprise that people have come to believe that flying modern jets is equally effortless.

    So, please, I am seriously looking to be educated on this. What makes the modern commercial pilot different, perhaps superior to the desk jockey in Nevada other than having the lives of hundreds of passengers on their hands. The AF guys make it look very easy and do give the impression that modern flying is just a simple matter of pushing a joy-stick with the computer doing everything including overriding any foolish inputs. I know it must be more than this, but as an outside observer, that is the impression one gets.

    • Patrick says:

      >> What makes the modern commercial pilot different, perhaps superior to the desk jockey in Nevada other than having the lives of hundreds of passengers on their hands. << I know it's tempting for the layperson to equate the two, but there's almost no comparison between a military drone and a commercial jetliner. They are vastly different.

    • Tim says:

      The drone pilots also have the lives of people in their hands. Although it’s kind of different because they kill them on purpose.

  21. JimBob says:

    I just flew to and from France in an AF A380. As seasoned a traveler as I am, I was still in awe of all that metal in the air, holding all those people, being flown with precision and complete (as it turned out) safety. That is one huge device, and I have to wonder whether the folks who think commercial aviation has become “automated” are mixing up their airplanes. To wit, if you put a modern set of avionics, autopilots, landing and warning systems and various redundancies on a Piper Cub, you would have a pretty much hands-off airplane. But I know g.d. well that the frogs flying that monster were not playing cards during take-off and landing, and were very much on their toes should anything even give a hint of going wrong.

    That said, I think the commenter who pointed out the vast difference between how we develop pilots in the U.S., where there is a healthy general aviation sector in which to garner thousands of hours of stick time before moving up to bigger stuff, is far more likely to produce resourceful pilots than a system that trains transport pilots from scratch.

    Next time, I’ll fly United.

    • Anonymous says:

      Langewiesche: “Pilots have been relegated to mundane roles as system managers, expected to monitor the computers and sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare even to failure.”

      Except for the U.S. & Canada, and a few airlines in Europe (and only a few) this is exactly how pilots are being trained today. The vast majority of new airline pilots do not know how to fly, they only know how to be systems managers. Stick and rudder, seat of the pants, manual flying, call it what you will: In a couple of decades when our generation retires, flying an airliner the way we know and do, will be a lost art.

  22. Anonymous says:

    “that preposterous, yet so sure-sounding caricature of a pilot’s job: “…sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare even to failure.” Maybe I’m fixating too strongly on such a small passage”

    I, for one, do not think you are fixating too strongly on that comment.

    At minimum it was a gratuitous insult to those in your profession.

    Even worse, it is that kind of bogus remark that pollutes the conversation and becomes a “fact” in the minds of laymen – call it the “everyone knows” syndrome.

    This can only be bad for pilots, crew, the airlines, the travelling public, and future air travel safety.

    No progress can be made, in any field, when the underlying facts are distorted. Where there is error, it must be exposed – for the good of all.

  23. Patrick,
    You and I see eye to eye on many issues. Although I agree again on the majority of what you say, I also agree with the Vanity Fair article.
    Piloting skills of the steam gauge ara has been replaced by FMC management.
    Fly by wire jetliners are much safer because of added technology but lack stick and rudder input from seasoned old pilots of Rope Start 747 days.
    I don’t just say that because I am one of those old pilots. As you know I still fly a very shophisticaked glass cockpit for food. I do notice I can’t do as good because my auto pilot does better. You may blame this degradation to my age which is partially true, but I blame it on complacency.
    On a couple of occasions when automation failed I did OK by reverting back to stick and rudders.
    I’m not sure I can say that about the new generation of pilots, “The Chidren of Magenta!”
    Captain Ross Aimer
    UAL Ret.
    Aero Consulting Experts
    Los Angeles

    • BTW, keep the word “fucking”. It is a realistic cockpit lingo.
      The Microsoft Flight Simulator Pilots may have an issue with that!

      • Thomas K says:

        As a Flight Simulator Captain – I fly the Level D B-767 for one of the largest Virtual Airlines on the net, I have absolutely no issue with a well-placed F-bomb, provided it is used surgically and sparingly.

    • Patrick says:

      Some people seem to be missing my point, which is maybe by fault in the way I worded this post. I am not arguing the author’s main contention, about the degradation of old-school, hands-on flying skills. What angered me about the story was his characterization of the job — the suggestion that it’s easy and involves little more than pushing a few buttons. I guess what I’m trying to say is that flying planes is still pretty complicated and challenging even for the WORST pilots.

      Maybe I’ll add something to this effect to the blog post itself.

  24. RenaissanceLady says:

    Your point here is perfect:
    “The argument in those days was that cockpits needed to become more automated, to engineer out this tendency. The unintended consequence of that approach, we’re now told, is that pilots can’t fly anymore. So what’s the fix? Yet more automation, or less? Or is that the wrong question entirely?”

    If anyone could push a few buttons and fly a plane, we’d all be pilots. That seems to be lost on those who assume all of the thousands of hours of classes, simulations and practice by those who become pilots is nothing more than wasted time as everything is automated. Then, when a disaster occurs, we can blame it upon the pilots lacking the necessary skills to fly a plane — which means we need even more automation to save us from unskilled pilots.

    I find this absolutely baffling.

    It seems (to me, at least) that more aspects of flying are becoming automated, yet I see no evidence of this leading to a general lack of piloting skill. Maybe I’m wrong, but I would therefore assume both skill and solid automation are necessary for a safe flight. Everything else just seems like a blame game in the attempt to find simple answers for complex discussions.

    • Patrick says:

      >> I would therefore assume both skill and solid automation are necessary for a safe flight. << Exactly. And in fact, contrary to what almost everybody seems to think, the levels of automation haven't really increased since the 1970s or so, with the advent of the FMS and autothrottle systems. When it comes right down to it, the newest 787 isn't really any more "automatic" than the 757, the A300, or the early 747s were, four decades ago.

      • Bill Palmer says:


        If you had said “the 1980’s or so” I would agree. Since the 757 was introduced (first flight 1982), automation has gotten a little better – but incrementally. The largest improvements have been in the display of information, the introduction of electronic checklists, etc., rather than the basic autopilot functions.
        However, to contrast today’s automation with say a DC-10 (first flight August 29, 1970) – there’s quite a difference. On that airplane, the autopilot was an analog computer. Many of the functions were a nice idea, but required closer monitoring and more intervention. The internal workings of the system were much simpler, predictions and flight planning functions were only rudimentary—when they existed at all. With the introduction of the A320 (first flight 1987), and the fly-by-wire aircraft that followed the flight controls themselves took on a whole other level of integration. Indeed, Airbus changed the very nature of the pilot controls to telling the airplane “what to do” not “how to do it.” And THAT more so than the non-interlinking of the sidesticks, is what needs to be understoood.

  25. SGM says:

    I installed Microsoft Flight Simulator on my first pc back in the mid-1980’s. Killed myself within minutes every time. Since then I have never thought that actually flying a real plane could be easy.

  26. chris conklin says:

    My takeaway from the VF article was very different from yours. I believe the author (I refuse to try to correctly spell “Langewiesch over and over”) has done a very competent job of presenting the complexities of our profession to the outside world. Yes, it is less than perfect, but it is the best I have seen- possibly ever. The article has been a topic of conversation in our cockpits, and the praise is mostly universal. I had the “pleasure” of flying the AF447 scenario in the sim (B767, not ‘Bus): we made it, but it was a real workout. Ive discussed the situation with some Bus drivers, and they believe it would have been more difficult in the Bus than the Boeing.

    We all grasp with how to present the reality of our profession to the laymen- you as much as any. I think that L. has done a great job.

    Chris Conklin

    • Patrick says:

      I don’t know Chris. You can’t possibly be comfortable with his description of the job as keeping your hands off the controls and pushing a few buttons. That’s just a disgusting way of describing it.

  27. Jeff O'Byrne says:

    Nobody writing about the AF447 incident, and especially not Mr. Langewiesche, has looked past the BEA/AF/Airbus “Blame the dead guys” approach to the investigation. That said, there is considerable material in the Final BEA report to indict both AF and Airbus. e.g., the flight management system knew the attitude of the aircraft throughout the flight but never displayed this to the flight crew. I have seen no reference to any independent indication of attitude in the cockpit of that A330. In any Boeing an independent attitude indicator is right at the top of the panel. In prior events when the pitot tubes were blocked, if the crew had independent attitude reference they flew out of the event safely. Second, why did the flight not divert 100 miles west to miss the clearly visible, massive ITCZ thunderstorm? Other company’s flights did so that night. Did AF restrict that option to save fuel? Given the lack of attention to these issues and the cacophonous atmosphere in the cockpit as the FMS cycled on and off, it is truly a miscarriage of justice to ‘blame the dead guys’.

    • Precisely. Anyone concerned with safety in any enterprise should remember Prof. Nancy Leveson’s maxim: “Human error is a symptom, not a cause.” Blaming the dead guys is just a way of keeping people from looking deep into your operation.

    • Duane says:

      “Blame the dead guys” is entirely appropriate when the principal, controlling blame for the accident was improper flight control inputs by the pilot flying … coupled with failure of the pilot in command to assess the PF’s inappropriate inputs (just look at his damned hand continuously pulling the sidestick rearward to its limits!) and declare that he “has the airplane”.

      Yes, there were other contributing factors, all of which were included in the final BEA report. What the author of the Vanity Fair article did quite well, however, was to recount (via CVR transcripts and FDR data) all of the attempts by the PF to climb the airplane to where he apparently thought there was safety above the clouds. His mindset in obsessing over a desired climb clearly was a major psychological factor leading to his incompetent control manipulations. His every action thereafter was to pitch the aircraft higher and higher til it stalled, and (except for momentary lapses when he obeyed the PIC to lower the nose), continue to do so until the aircraft crashed into the ocean. Reading this account, the PF was obviously overcome by fear and ignorance.

      There is no way not to blame this accident mostly upon the incompetent actions of the PF, as aided and abetted by the incompetent failure of the PIC (and later on the captain) to assert control over the aircraft from the panicked PF. Additionally, the captain should never have taken relief from an under-qualified co-pilot as the aircraft approached a known weather system, nor shown up at his job with no reasonable rest in the prior 24 hours.

      Sure, there are engineering design improvements that could lessen the impacts of the PF’s incompetent flying, such as better visual cues from the FD and stall alarm, and linked sidesticks … but any such improvements in visual and mechanical clues can be ignored by the flight crew just as easily as the flight crew ignored the attitude indicators (which never malfunctioned, and clearly displayed the grossly high pitch angles), and the altimeter, and the GPS groundspeed, and the physical buffeting from the stall, and so forth.

      No amount of re-engineering can make any aircraft entirely idiot-proof. The reality is, if the PF had simply kept his hands off the controls for the minute or so that the airspeed indicators were out of action, the airplane would have continued flying on and no accident would have resulted.

      Defending the dead guys in the cockpit, on the other hand, is entirely unproductive, indeed anti-productive. That AF and the airline industry need to do a better job of qualifying and re-training their pilots to deal more competently with system failures at high altitudes goes without saying. Yet no amount of training or engineering redesign can make up for a fearful and temporarily panicked, indeed effectively insane PF who does not understand what the airplane is actually doing, and who furthermore refuses to give up control of the aircraft to his PIC. Nor can any such improvements make up for an under-qualified PIC who refuses to affirmatively assert control over the airplane when he has a fearful, indeed crazy PF in the right seat.

      • Jeff O'Byrne says:

        The real picture is a bit more complex than Duane states . As stated on page 190 of the English translation of the BEA Final Report (BEA): “Until the end of the flight, the angle of attack values changed successively from valid to invalid. Each time that at least one value became valid again, the stall warning re-triggered and each time the angle of attack values were invalid, the warning stopped. Several nose-down inputs caused a drop in the pitch attitude and the angle of attack, whose values then became valid, such that a clear nose-down input resulted in the triggering of the stall warning. It appears that the PF reacted, on at least two occasions, with a nose-up input, whose consequences were an increase in angle of attack, a drop in measured speed and consequently stopping the stall warning. Until the end of the flight, no valid angle of attack value was less than 35°.” In addition, as the BEA states on page 198 “ˆThe aeroplane’s angle of attack is not directly displayed to the pilots.”
        In addition, BEA page 21-22 states “At 2 h 10 min 51, the stall warning triggered again, in a continuous manner. The thrust levers were positioned in the TO/GA detent and the PF made nose-up inputs. The recorded angle of attack, of around 6 degrees at the triggering of the stall warning, continued to increase. The trimmable horizontal stabilizer (THS) began a nose-up movement and moved from 3 to 13 degrees pitch-up in about 1 minute and remained in the latter position until the end of the flight.”
        In other words, the pilot flying had a flight director giving him a stall warning when he pushed the nose down and silencing the stall warning when he pulled the nose up. Further, something told the trimmable horizontal stablilizer to go to full up and stay there. The crew wasn’t the only confused entity that night. Even the flight management system was having problems!
        Blaming the ‘…dead guys…’ is wrong but more importantly it leads in the wrong direction. Computerization is valid only within the imagination of the programmer. It is the responsibility of the programmer to ensure that the system degrades gracefully when that imagination is exceeded. In this context that means going to one state and staying there. For the entire system, that means an independent attitude indication. The A330 FMS gave conflicting indications and did not provide an independent attitude indicator. Without the ultimate backup, a visible horizon, the incident was not recoverable.

    • Bill Palmer says:

      >>Nobody writing about the AF447 incident, and especially not Mr. Langewiesche, has looked past the BEA/AF/Airbus “Blame the dead guys” approach to the investigation. <>That said, there is considerable material in the Final BEA report to indict both AF and Airbus. e.g., the flight management system knew the attitude of the aircraft throughout the flight but never displayed this to the flight crew. I have seen no reference to any independent indication of attitude in the cockpit of that A330. <>In any Boeing an independent attitude indicator is right at the top of the panel. In prior events when the pitot tubes were blocked, if the crew had independent attitude reference they flew out of the event safely.<>Second, why did the flight not divert 100 miles west to miss the clearly visible, massive ITCZ thunderstorm? Other company’s flights did so that night. Did AF restrict that option to save fuel? Given the lack of attention to these issues and the cacophonous atmosphere in the cockpit as the FMS cycled on and off, it is truly a miscarriage of justice to ‘blame the dead guys’.<<

      Other flights, including other AF flights diverted around those storms. The captain just didn't seem interested to do so. I believe there is also evidence that the radar was not being operated in a way to display them properly. When FO Robert returned from his break minutes before the loss of airspeed he adjusted the radar display and suggested a deviation to the left, albeit only a few degrees left at that point, but by then they were moments away from entering the storm.

    • Bill Palmer says:

      >>Nobody writing about the AF447 incident, and especially not Mr. Langewiesche, has looked past the BEA/AF/Airbus “Blame the dead guys” approach to the investigation. <<

      Not True!
      For a much more in-depth explanation of what happened and why, from crew rest to communication breakdown, the weather, flight controls and of course analysis of the pitot tube issue, as well as the pilot's experience and training, please see my book "Understanding Air France 447." It's NOT a 'blame the dead guys' approach. The core information and analysis was written for other Airbus pilots, but there's plenty of background that any aviation enthusiast can grasp the technical details.
      Included is a CVR/FDR combined timeline, compete analysis of the ACARS data – more than the BEA report will tell you, what was required to recover as the scenario progressed, how the Airbus fly-by-wire controls work and the part that they played. How the flight director-giving erroneous information probably heavily influenced the pilot flying's failure to recover to level flight. I take you step by step through the accident, show you the flight recorder data traces and explain what they show.
      I'll also tell you why an angle of attack indicator isn't a cure-all, discuss pilot training, and what needs to happen in the near future in regards to it.
      If you're at all interested in this accident, I think you'll find it a worthwhile read.

  28. UncleStu says:

    I expect much better from Langewiesche.

    Considering that he knew you well enough to write a blurb for you, a phone call wouldn’t have hurt. I wonder if he was just being lazy in his research.

    Regardless of the reason, there is no excuse.

  29. Jay Becker says:

    I hope that you are feeling better, and the act of getting sick, especially out of town, is unpleasant, and difficult to avoid. It is hard to be careful even at home.
    You made a good response to the Vanity Fair article. As a fairly frequent air traveler (every other month), I have the greatest respect for the crews that run the system. I have less respect for the TSA screeners at the airports, and somewhat less than that for those who run the “food courts”.
    Thank you for the writings you produce; they are an important service to the flying public.

  30. Anil Pillai says:

    At one point he writes of pilots: “All of them think they are better than they are.”

    What a stupid comment that is going to lose him a lot of respect. It is almost like he is not aware of the irony – the same could be said of writers.

    And of course, his visions of the future of flight in a kingdom far far away:

    “Pilots have been relegated to mundane roles as system managers, expected to monitor the computers and sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare even to failure.”

  31. Buff Crone says:

    Send a letter to the editor of Vanity Fair, as an active pilot and author of Cockpit Confidential. At a minimum it will boost book sales!

  32. Peter Fulton Foss says:

    Dear Mr. Smith:

    I was offended on the one hand and surprised on the other that you felt it necessary to use the word “fucking”!

    You have an excellent command of the English language, which made your choice of this adjective even more surprising.

  33. Ellen says:

    Great rebuttal – thanks for your thoughts. The Letters to the Editor in this month’s Vanity Fair are loaded with responses to the Langewiesche article. Many of them are knee-jerk reactions, but one writer makes the same point you do about the Airbus design issues and the role they might have played.

  34. Roger Wolff says:

    As far as I’m concerned, the AF447 flight was flown into the ocean by the pilots. From the official report the captain recognized the stall and told the flying pilot to PUSH the stick not pull. He did so for a few moments and then proceeded to go against the captains orders and started pulling again.

  35. Tim says:

    Patrick (and others),

    You might be interested in this event in Boston on December 8. Your buddy R. John Hansman of MIT will be discussing cockpit automation before a screening of Airplane.

    Be sure and ask good questions.

  36. A historical shift IS occurring. Airmanship is being re-defined. Transition is never easy and it is never over.

  37. Mark R. says:

    If you don’t read the newspaper you’re uninformed.
    If you do read the newspaper you’re misinformed.
    — Mark Twain

    Vanity Fair is slick but not a place I look for accurate “information.

  38. wonkazoo says:

    Hi Patrick,

    You and I have corresponded off and on over the years- I hope this finds you doing well.

    Normally I am squarely in your court, but on your review of Langewiesche’s article I think you are taking it a bit too personally.

    A little more than a year ago a crew of three flew a visual approach right through the glideslope because of a failure to understand what was automated and what wasn’t in their airplane, to the detriment of the 777 and several lives. AF 447 takes this to an altogether different plane- and in my mind at least it isn’t just the automation- it’s the physical disconnect between the two control sticks.

    Essentially the flying pilot pulled up in a panic and no one in the cockpit was able to recover from that initial reaction- and the confusion caused by the automation/disconnected sticks was the secondary/tertiary cause of the disaster. As a commercial pilot I know well the skills necessary to safely and capably fly any aircraft, but to assert as you sort of do here that an additional factor was that Bonin was essentially a fuckup and is an anomaly I think misses the greater point.

    The notion that Bonin (and the rest of the crew) were an anomaly, and that there isn’t a greater issue facing modern day aviation is clearly not correct, and recent history has shown that the greatest threat to modern day airliners appears to be the dulling of the most basic skill of all: FTFA.

    Perhaps the author could have done a better job of recognizing that even the most unskilled or talentless of commercial or ATP certificate holders have at least a basic level of skills and ability, and certainly advocating for pilotless airplanes is utterly foolish, at least for the foreseeable future, but otherwise I thought his article did a fantastic job of illuminating how utterly inane and silly the entire debacle really was.

    I mean think about it- quite literally things would have turned out better if both pilots at the controls had been asleep (and stayed that way) when the airspeed indicators went offline.

    When an unconscious pilot performs better than an awake one something has gone wrong- seriously wrong, and in this case the mechanical failure pales in comparison to the profundity of the failure of basic airmanship displayed by the flying pilot. Does that mean that all pilots in today’s world can’t demonstrate even the most basic of skills?? Nope. But it is indicative of a glaring and amazingly profound issue- one which has likely been brought about by the nature of the modern-day airliner cockpit.

    Anyway keep up the good work as always, and as always: Keep it rubber side down when landing!!


    • cthulhu says:

      The disconnect between the sticks is a serious Airbus design flaw (there are others too; in general, their redundancy management is atrocious – I speak as one with a a lot of experience in the field). There has been a LOT of work over the last 30 years on design of what are called in the industry “control inceptors”, and a near universal finding is that the inceptors in a dual station flight deck must be linked together and must be back driven by the autopilot so that pilots know, when assuming manual control or taking over control from the other side of the flight deck, exactly what the controls are doing.

  39. Eirik says:

    I think it was a very interesting article, and something that should be taken seriously. No matter what kind of profession, when you are used to the machine doing the job 95% of the time (although with human input), it may come as an unpleasant surprise when, all of a sudden, you are in command.

    Im not saying pilots dont know how to fly (by all means), and I think the author takes that argument way too far, but getting frequent flying training and going back to basics a few times each year is never a bad idea.
    By that I mean flying real planes, not a simulator in safe conditions.

    Having said that, the biggest irony in the article is the authors own fact; one fatal accident for every five million departures.
    Which leads to the following question; what is the problem?

    Of course, every accident is tragic and one too many, but there will always be accidents, both in the air, on the road and at sea. It is simply inevitable. Since accidents are so rare these days, the whole world is watching when it happens and sometimes the reason for the accident is blown out of proportion.

    Just because the crew on flight 447 made fatal decisions which lead to the crash, does not mean there is a pilot problem.

    The statistics speaks for itself, and every day there are pilots around the world making the correct decision which makes flying very safe. Unfortunately, that is not newsworthy and we all take it for granted.

    • Tom says:

      The full report is avail online and makes for fascinating reading. Here:

      One of the things that stuck out in my mind was the overwhelming number of visual and aural alarms all going off at the same time in AF447’s cockpit, leaving the pilots to figure out what each subsequent alarm was telling them and exactly how to get the whole cacophonous mess to shut up so the pilots could do what ALL pilots are trained to do when facing a crisis, which is FLY THE PLANE!

      Commercial pilots today are all trained in high fidelity simulators. In many cases, they get signed off to fly a certain plane type before actually sitting in the real thing. Often their first flight in the real thing is also their first flight carrying passengers (comforting thought). Similarly, they train in the same simulators on how to recover from all sorts of emergency situations–loss of power, radios, weather situations, stuck landing gear. Though useful, such exercises are inherently limited because the pilot walks into the machine knowing he or she is going to have to deal with some sort of in-flight situation. Real life isn’t so accommodating.

      As noted by Langewiesch, Bonin was reacting to the alarms, not the plane. He held the plane at a nearly impossible 13* stall angle because the plane’s aural warning shut off at that angle. When he lowered the nose it reactivated, and like Pavlov’s proverbial dog, he instinctively returned the stick to a place where the alarm went silent.

      Equally amazing is how much time these pilots had to resolve the situation. How many minutes passed and how horribly they worked in concert to try and recover the plane. One has to wonder if Bonin, Robert or Dupuis had been alone in the cockpit and left to their own skills and resourcefulness to recover the plane that things might have worked out much differently.

      • Eirik says:

        Thanks for sharing the link! Interesting read for sure, although some parts are too complicated for me 🙂

        On a side note, in case you liked the Vanity Fair piece, if for no other reason than the read itself and not the technical debate, I can recommend this one;

      • Duane says:

        Tom – one of the BEA findings was that the airline provided little to no training in its recurrent simulator training programs for high altitude systems failures. That certainly contributed to the flight crew’s inability to manage a relatively simple loss of airspeeds scenario.

        The so-called “cacophony” of alarms, however, is actually easy enough to ignore for a highly trained pilot flying who knows how to obtain and focus on only the information necessary to fly the aircraft. For pilots who are not well trained in systems failures, or who are otherwise incapacitated by fear, however, the alarms simply add to the confusion and panic of not knowing what you’re doing at the controls.

        As to Bonin “flying the alarms” rather than by the flight instruments, if accurate that in itself tells us he did not really know how to fly the airplane. Or at least, he didn’t know how to fly in his state of mind during the accident. He pulled the sidestick all or nearly all the way aft to its stops throughout most of the time that he was attempting to manually control the aircraft. There is no known flight regime for any fixed wing aircraft in which doing that would be appropriate. It seems pretty clear then that the PF was not operating according to any known flight procedure, but rather was instead virtually incapacitated by fear and panic. Fear and panic, once it takes hold, can render anyone in its grip temporarily insane.

  40. Alan says:

    Oh, boy. When I first opened this article I thought the topic was about Wolfgang Langewiesche author of the book Stick And Rudder I learned to fly from many decades ago. It was pretty much required reading in our flying club. It took me another moment to remember that Wolfgang died about a decade ago.

    So I just looked it up and learned that it is Wolfgang’s son that we are talking about.

    • Rod says:

      Until I encountered William, I’d never heard of Wolfgang. But I know what you mean. Richard Collins and Sparky Imeson are the writers who helped me learn to fly. As they say, the cockpit of a single-engine plane is a crappy classroom. But a nice cosy book — that’s when you put two and two together.

  41. georgie says:

    it is always important to have 360 degrees of information before laying blame, fault, when it has resulted in loss of life. There isn’t an aviation disaster that cannot be taken back a step or 2 further to analyze what happened. Thank you for this post

  42. Duane says:

    As a private pilot, I can understand the anger of professional airline transport pilots at Langewiesche’s flippancy concerning pilots skills and attitudes and value to the air transport system (he apparently thinks it is next to nil). Throughout the article he repeatedly kept hammering professional pilots, so his agenda was clearly anti-pilot (despite his own aviation background). But he still misses the mark.

    The problem with AF447, as well as Asiana 214, was the fact that those flights were conducted by bad pilots. It is impossible eliminate all bad pilots – the hiring and certification of pilots is itself a human endeavor that is subject to human error. That occasionally bad pilots cause accidents only provides motivation to make continual improvement in the training and qualification of pilots … it does not tell us to eliminate human pilots, which cannot be done safely. Who would want to fly on a commercial airplane where the guy or gal in command does not have their own life on the line, i.e., “no skin in the game”? Not me certainly.

    Bonin was a bad pilot who, essentially, through his fear of the weather and of the airplane, went crazy at the controls and manipulated them in a way that even a first hour student private pilot is taught never to do. His relief PIC, also inexperienced, should never have been PIC, and his skills were not up to the task. His captain was negligent as well, as was the airline (Air France) that put this collection of incompetents together in a cockpit.

    As with most aviation accidents, AF447 illustrates the need for a number of improvements that can and ought to be made, including, but not limited to:

    1) Much more rigorous initial and recurrent simulator training on high altitude non-standard operations (system failures, unusual attitudes, etc.). Automation itself does not eliminate or prevent training – only training not performed eliminates training.

    2) Revise the stick control to allow an absolute override by the left seat. Someone has to be in final command when a right seat pilot flying goes nuts. Link the sticks electromechanically too – it’s an obvious need.

    3) Make the necessary improvements to the airspeed indicators – go with fully electronic, laser-based sensors that cannot ice up (the technology exists and has already been patented).

    4) Automation is here to stay … pilots and those who employ them must nevertheless make sure all pilots can manage system failures.

    • Wol says:

      >> His captain was negligent as well, as was the airline (Air France) that put this collection of incompetents together in a cockpit.<<

      That's not entirely fair, although there's quite a lot in what you say.

      What seems to be ignored most of the time is the contribution of the Airbus philosophy with regard to the non-existent visual feedback of the controls, especially (in this case) of the sidestick.

      The captain re-entered the flight deck in the middle of a chaotic situation and probably should have got into the left hand seat immediately(easy to say after the event.) However, he didn't – and was completely in the dark as to what control input was being applied. It's inconceivable to me that had he been able to see a control column being manipulated in the way the sidestick was that he couldn't have given instructions from the back of the seats and regained control.

      I admit I've never flown a sidestick aircraft, but having sat in jumpseats watching them being driven by hand I can't say the lack of feedback to the non-handling pilot is anything but disturbing.

      • Duane says:

        The only appropriate action for the PIC in this instance would have been a loud and unmistakable statement to the flight deck of “I have the airplane”.

        The PIC never did that, even though the FDR and CVR shows clearly that he attempted at various times to control the aircraft. Even the lowliest, least experienced student pilots and their time-building flight instructors know what to do when the PIC makes such a declaration of sole authority to manipulate the controls. Surely a qualified air transport flight deck crew would do that, right? Apparently not on this flight.

        As to whether the PIC could discern the incorrect manipulations of the sidestick by the PF (the FDR shows clearly the PF had the stick pulled all or nearly all the way to its aft stop during nearly the entire flight accident sequence – which should have been noticeably visible), if the PIC had made the simple, standardized declaration of control, then that should have taken care of matters and recovered the aircraft. Regardless of what the PIC could see or feel, or not, in the controls.

        But he didn’t.

        It is highly likely that that the relief PIC’s own inexperience and lack of recency, as described in the article, may have discouraged him from seizing the controls authoritatively as required. The fault for that situation lies both with the relief PIC as well as with the captain, who knowingly took relief from an inexperienced and under-qualified pilot, and who knowingly deprived himself of sleep prior to the flight thus prompting his early need for relief due to fatigue. The airline is also at fault for not providing an experienced relief captain for the flight, instead of two inexperienced first officers.

  43. Dennis Boykin says:


    I’m certainly not in complete agreement with your assessment of the article, and here’s one where you got it wrong: . . . . on average the knowledge base has become very thin.” He cites “nearly universal agreement of this,” but I’ll tell you it simply isn’t true, and there’s a reason why, in this country, the major commercial airlines expect a new-hire pilot to have somewhere around 6,000 hours of flight time before they’ll take his or her resume seriously. . . . .

    “In this country”? These were French pilots flying under a completely different business model & aviation industry paradigm – ab initio training, no GA industry to speak of, and limited exposure to anything in aviation outside of what Air France required of them. The North American, Australian, and (somewhat) English model of pilots coming up thru the industry (as flight instructors and commercial pilots on smaller equipment) is completely foreign to the other European carriers, as well as most of the rest of the world. The ab initio program puts young pilots into complex machines far, far sooner in the rest of the world than it does in our system.

    Dark outside, IMC, whatever – none of that matters – the real problem was a lack of experience in attitude flying. There was one primary cause, in my opinion, and one contributory: The 1st Officer failed to fly the plane, and the Airbus design contributed to that problem. Yes, he missed the supporting problem, but the primary remains clear.

  44. Todd says:

    Langewiesche needs to consider how many times good pilots save the day when similar events occur and simply go unreported because disaster was averted. Put him in a simulator and see how he does.

  45. Keith McLellan says:

    Patrick is right on target here with this article. I am one of those who believes that generally speaking, basic flying skills have deteriorated over the years, at least as evidenced by my own personal observation of many pilots I have flown with over the past 10 years. Having said that, though, I agree with Patrick that this is in part due to the fact that pilots today are spending as much time learning how to use modern avionics and all the automation that comes with it, and having to do so almost from the very beginning of their earliest flight training.

    Whereas I learned to fly airplanes on “steam gauges” and learned to fly without the use of many instruments that most would take for granted, today’s new pilot is using highly sophisticated avionics and instrument displays, often from the very first lesson they ever take. Today’s small, single engine aircraft are often equipped with “glass” that would have been the envy of most airline pilots only 20 years ago.

    The problem this causes is that the new pilot spends more of his or her time trying to learn how to use the avionics and automation, rather than focus on basic flying skills and the physics of flight. I’m always amazed at how much many of the newest pilots I have met recently did not know about basic aerodynamics, or knowing the right answers to my questions, could not demonstrate in the cockpit what I consider to be the most basic and requisite skills necessary to safely be in command of an aircraft.

    I blame this situation on the instructors and flight schools that are instructing the pilots. The use of modern avionics and automation should be secondary and taught as a follow-on to the more important (in my opinion) basic knowledge of aerodynamics, aircraft handling and hands-on flying skills.

    I can give many examples of pilots I have flown with in jets the past few years who have exhibited a lack of such basic flying skills and proficiency. I have often seen a surprising lack of skill or ability to control altitude, airspeed, rate of climb or descent, glide path on an approach, or set up and fly a simple VFR pattern around an airport to a visual approach and landing.

    Learning how to program and use the avionics and automation in a new cockpit should be something you add to your skill sets as a pilot. It should NOT replace the very basic skills of flying the aircraft with only a minimum of instrument indications. How many today really understand the “old” concept of “pitch and power equals airspeed”, or have the skill to use it when needed? I have flown a jet for several hours that had lost all airspeed and altitude instruments (which included vertical speed instruments). I’ve always been taught to know what target power settings are normal for the various configurations of my aircraft, especially with regard to descent, approach and landing, so successfully getting the aircraft back on the ground seemed simple and almost routine, and I don’t think my blood pressure or heart rate went up by very much.

    Another problem with many modern day accidents is that I am seeing a lower overall average of hours and experience on the part of the pilots involved. Hours is one thing, but by experience I mean the amount of flying, as pilot-in-command, of different aircraft and different types of flying. Many newer pilots, and often times many airline pilots as well, do not have as wide a base of experience to draw from as pilots from my generation have.

    I am typical of many pilots of my age, at least in America. I had flown several models of each of the popular small single and twin engine aircraft that were common in the 70’s and 80’s (Cessna, Piper, Beechcraft, Grumman, North American, Bellanca, and Mooney). By the time I finally got a first interview with a commuter airline, 5 years after I first started taking flying lessons, I had flown 27 different models of aircraft and had over 3,500 hours total flight time (of which roughly 3200 hours were as pilot-in-command, and all but the first 500 hours were flying for small “mom and pop” charter and air tour companies).

    That commuter airline job had me often flying as many as 5 legs in a day (my record is 9). It was and still is normal to see solid IMC and approaches down to minimums much of the time in certain areas. As a First Officer, I got to fly half the legs each day of any given trip, and at all times under the supervision of many different Captains with a lot more experience than I had. I learned a lot from all of them, and many of them were happy to pass on to me what they had learned over the years. Although I turned away from an airline career and became a corporate pilot instead, I am happy that most of the American flag carriers are crewed with Captains who came up through the ranks like I did, and have similarly deep levels of experience and skills to draw on.

    I have watched the airline industry outside the USA for years, and I have always thought that the ab-initio training programs of many foreign carriers is not a very good way to train pilots. It puts too many pilots in the cockpit of large airplanes with little in the way of the real world, hands-on experience I’ve described above. I think the Air France 447 accident is an example of this, and proves that pilots need to be and are much more than just “system managers” or “button pushers”.

  46. I’m not a pilot, but I’ve worked on digital flight systems since the first one (the JA37B autopilot) and I suggest Mr. Langewiesche move from the general to the specific and perform a simple exercise: go through the cockpit design described in that accident report and for each input and display, ask if the way it works enhances or detracts from the pilot’s situational awareness. My intuitive reaction upon reading the report was detract beats enhance 10 to 1. I may be misrembering this (happens a lot a my age) but I seem to recall that the cockpit designers decided that an angle of attack indicator was unnecessary.

    There is automation and there is automation. Good automation provides more situational awareness than the old days of hanging out the window. Bad automation provides a virtual cocoon, devoid of feed back, that converts a pilot into a spectator.

  47. Patrick Wright says:

    People think airliners are easy to fly because the airplane games on their home computers and iPhones are easy to fly.

    • JohnLM says:

      I can attest to that fact. I’m a regular sim flyer and I was able to spend time in a 737 simulator and it was like night and day. While having knowledge of the systems and cockpit layout the actual flying characteristics are nothing alike. Part of it is because everything is spaced out across a large space instead inches apart on a computer sim making it much more difficult to manage and in the same vein the professional simulator gives you the feeling of having the rest of the plane attached to the flight deck which is disorientating to non pilots such as myself. I kept the plane aloft and landed (with lots of help from the instructor) but if there had actually been passengers they would have been puke all over the cabin, it was very hard to keep level even though I instictiveley knew how to work the trim tab. And if it’s anything other then a clear day forget about, as soon as you take your eyes away from the instruments and try to fly by sight you become instantly disorientated. I was over Hong Kong with 1000 ft ceiling blanketing the city and on approach to Kai Tak. In a mere seconds I got flustered and over saturated with tasks and basically nose dived into the airport about 20 degrees off the centerline just to get below the clouds. I made the landing but as the instructor said after, “we should probably wait for the authorities to show up before leaving the cockpit.

    • Thomas K says:

      Ever fly the Level D 767 ( with real-world weather, and ATC through VATSIM? Give that a try then come back here and talk to us about how easy it is to fly simulated airliners. Not the real-world heavy metal of course; but MUCH more than a computer game if done properly.

  48. Chris says:

    I quit paying attention to Langewiesche — who IMO is largely trading, at least in the aviation journalism business, on his family name, rather than any particular expertise in aviation — around the time he essentially made up the article five years ago, also for Vanity Fair, about the midair collision over the Amazon between an Embraer Legacy and a 737:

  49. Ma Zhenguo says:

    I know this is so obvious that it hardly needs mentioning, but I will still point out that this accident would not have happened if the sky weren’t pitch black, in my opinion. In broad daylight, the pilots would easily realise that their attitude was wrong enough to cause a stall, and probably correct the input accordingly. So there you have a long series of factors ultimately leading to the crash: (1) pilot inexperience (2) perhaps a culture of entitlement of the captains at AF (3) faulty pitot tubes (4) bad weather, or the fear of bad weather (5) a sleepy captain who hadn’t rested enough and wasn’t in the cockpit when he probably should have been (6) the independent controls of the A330, because of which the pilot on the left seat didn’t realise what Bonin was doing (pulling!!) (7) the lack of direct vision, i.e. outside it was pitch black (8) bad CRM (9) anything else?

  50. JamesP says:

    I use the cruise control on my car regularly. Granted, all it does is maintain speed, but I still have to get the car from my garage, across town, merged into traffic, and settled into a lane before I can even think of engaging it. It’s helpful and makes driving less fatiguing, but it doesn’t mean I don’t need to know how to drive.

    I once saw a color-coded graphic of the approach and departure routes in the San Francisco Bay Area. It covered SFO, Oakland, and San Jose airports as well as Moffett Field. Holy (bleep)! I wouldn’t think autopilot could handle that on its own.

    Ah, I just did a search and found it:

    I would imagine that keeps *both* pilots pretty busy.

  51. JuliaZ says:

    Patrick, if I’ve learned anything from reading you over the years, it’s that very few accidents are the result of a singular event… take-off and landing problems like Concorde suffering a fuel tank rupture from shrapnel on the ground and then crashing in Paris, microbursts, or bird strikes notwithstanding, usually there are two, three, or even half a dozen problems and errors chained together that combined, become insurmountable no matter what the pilots’ skill levels are.

    This crash seems no different. Multiple problems: exhaustion, poor communication, misinterpretation of instruments, insufficient warnings from the systems, and so on. Nobody will ever be able to say what could have broken the chain of problems to prevent the stall and resulting tragedy… the article makes several claims that made me roll my eyes. I read that article on my Kindle on AS SEA – DCA and my neighbor looked at me funny when I groaned out loud over this: “a different crowd is flying now, and although excellent pilots still work the job, on average the knowledge base has become very thin.” I might buy that criticism of minimum-wage regionals, but NOT on mainline metal. It takes too many years of ridiculous hours and poverty wages to get enough experience to work flights like Air France 447 for that accusation of “thin knowledge base” to be true.

    I am so sorry pilots don’t get proper respect for the multiple skillsets they have to develop and maintain. I get mad when people imply that pilots are “along for the ride, because the plane flies itself.” Yeah, and cruise control means your car drives itself too, right? Idiots.

    “Dr. Google” has similarly convinced many people that their doctors are idiots, and while it is true that someone graduates at the bottom of every class, most doctors care intensely about their work and have invested an incredible amount of time and money in their education.

    It is really disappointing to read a writer who clearly does have an effing clue arrive at such naive conclusions, conclusions that his own article doesn’t support. It’s like he argued with his editor, who wanted a different conclusion, and lost. I was sorry, ultimately, that I’d paid money to read it instead of reading the free Kindle/Prime book of the month.

    • bobbi says:

      > pilots don’t get proper respect for the multiple skillsets they have to develop

      There’s basically 2 skills:

      1. Feed the FRONT end of the bite-prone doggie that supervising the pilot who’s supervising the computer that flies the plane.

      2. Clean up after the BACK end of said canine.

      I bet somebody on AF 447 forgot to feed the dog…

    • UncleStu says:

      “It’s like he argued with his editor, who wanted a different conclusion, and lost.”

      If that is so, and Langewiesche knew the result would be that bad, he should have withdrawn and kept his integrity and reputation intact.

    • Bill Palmer says:

      When the 757 first came out, and I was instructing experienced airline pilots how to fly it, it was essentially a course on the automation. There wasn’t much hand flying involved – unlike the checkouts in the 727. Indeed, if they couldn’t hand fly a Boeing by then, they wouldn’t be there.
      That was a valid assumption for quite a while. Underlying hand flying skills were already there. They didn’t need to be taught, or really even checked, except for a few key maneuvers.

      A generation later, the same assumptions may no longer be true. Even a Cessna 172 now has a glass cockpit, including a flight director. When I flew for a commuter airline in the early 80’s we didn’t even have a autopilot. Every leg, every ILS, every non-precision approach to minimums: hand flown. Now there is a real possibility that a pilot may not have ever flown without one. Yet the training on the heavy metal has not quite evolved in parallel with this realization. The recent revision to FAR 121 training and simulator standards in the wake of Colgan 3407 and AF447 crashes will hopefully help in these areas.

  52. Dan Ullman says:

    If am not mistaken flight 447 was promoted as a mystery for awhile.

  53. David Magda says:

    I think the video “Children of the Magenta” (from 1997!) talks really well about piloting and automation:

    According to the talker, there are three levels of ‘automation’: manual (stick and rudder), autopilot, FMS/GPS.

    Sometimes, to reduce workload, one must actually go from a “higher” level (FMS, AP) to a “lower” level (FMS->{AP,manual}, AP->{manual}).

    (The video often gets posted and pulled because of copyright claims by American Airlines.)

  54. Josh says:

    Unfortunately, Patrick, since we live in the age of instant and accessible information via the internet, many professions which require years of training and experience are belittled as something that can be learned on the fly.
    I too am in medicine and cannot tell you how many uneducated folks tell me what their diagnosis is or what medicine they should be taking. But there IS still a role for proper training and experience and I doubt even the most sophisticated flight simulator devised will ever substitute for the real thing.

    • bobbi says:

      > many uneducated folks tell me what their diagnosis is or what medicine they should be taking.

      Half of them self-diagnose THE SAME DAY they read the latest arrival from the Reader’s Digest subscription.

      Patient: “Doctor, Doctor! I’ve got a rare and fatal disease just like it says in ‘The Digest’! What should I do?”

      Doctor: “Stop reading Reader’s Digest.”

  55. Captain Jerem Zwart says:

    No amount of automation can cope with complex, rapidly changing weather. At the end of the day, a pilot has to be trained to an acceptable level. The author assumes that all pilots are trained to the same minimum standard. That simply is not true.

  56. John LM says:

    While I thought the article was an interesting read, especially during the narration portion I found many of the conclusions to be premature and a little across the board. I was especially intrigued by the anxiety of Bonin and how that may have contributed to his initial reaction to climb. In all the earlier reports or from the released CVR transcript I didn’t pick up on his increasing anxiety while flying below the cloud deck. In his mind clear air was a scant 1000 feet above so when the autopilot disconnected it seems like his initial reaction was to finally get above the soup.

    I think the author was referring specifically to the new breed of Airbus pilots as systems managers rather then professional pilots as a whole and it was Bernard Ziegler himself who said he thought a concierge could fly his plane (although I’m sure tongue was planted in cheek).

    I’ve read a lot of material about this accident and there is definitely a pattern of Airbus pilots deriding Bonin as the main cause with his counterintuitive manipulation of the controls while the rest of the aviation community pointing to technical perimeters that could have prevented him from doing so. Airbus pilots point to the exemplary safety record of the fleet all without the need for dual input controls. Had the pilot been trained correctly or adhered to his training the outcome of this would merit not even a blip of attention. Supposedly the AOA gauge is optional but available on Airbuses and maybe it’s time it becomes standard.

    As with all accidents this was definitely a chain of events, from Dubois’s indifference to the soon to be replaced pitot tubes to the ditching of the CRM handbook and Bonin’s fatal actions. The first two had probably been in play for years. Dubois didn’t just pick this flight to loosen the proverbial tie around his neck, that kind systematic relaxed behavior devolves over years. The pitot tubes, while substandard, were not defective and had safely ushered millions of flights previous to this one. Bonin caused the plane to stall but had CRM been followed would the crisis been averted?

    It seems like to me the perfect void filled that cockpit at the worst possible time and just like 9/11 it’s something that will probably never happen again because it became the ultimate teaching moment.

    • Mike K says:

      As a surgeon, I have for many years referred to the “DC 10 cockpit syndrome” in the operating room when things happen too quickly to cope with all the decisions that have to be made. I was referring to the Chicago DC 10 that lost an engine on takeoff many years ago. Time in a simulator might come up with a solution but they were a hundred feet above the ground and had no time. I’ve been there in the OR a few times. Long ago, I was an aeronautical engineer but not a pilot.

  57. Rod says:

    “(Langewiesche’s) conclusion is a familiar and, to somebody who flies for a living, irritating trope: pilots have become so reliant on automation that we no longer know how to fly.”

    And yet his advocacy of Airbus’s fly-by-wire concept in his book on the subject basically promises that it’s meant, precisely, to save fallible pilots from themselves. So is he now contradicting himself?

    “(…) another factor (…) possibly crew fatigue”

    There has been precious little attention paid to the fact (as reported in the French media) that hotel staff and others in Rio thought the captain looked absolutely exhausted before departure.

    “I wonder if he’d make such a rude and cursory blanket statement about doctors or other professionals.”

    My guess is Yes.

    Langewiesche: “Pilots have been relegated to mundane roles as system managers, expected to monitor the computers and sometimes to enter data via keyboards, but to keep their hands off the controls, and to intervene only in the rare even to failure.”

    Well, his book on the subject implies that this is exactly how it OUGHT to be.
    I haven’t read the Vanity Fair piece, but it sounds like he’s backing away from his earlier stance.

    • Bill Palmer says:

      Fatigue, Indeed! There is a French TV show that reported a tour guide on a helicopter tour they took that afternoon also reported them to be tired. The captain had supposedly remarked he would sleep on the plane.

      By the time of the accident, it was after midnight Brazil time and even later in Paris——their body clocks perhaps somewhere in between. Upon returning to the cockpit 10 minutes before the loss of the autopilot, FO Robert reported that he got poor sleep on his break: “dozed a little.” Bonin apparently had looked tired enough that the captain offered him the opportunity to take a nap in his seat.

      One must also consider the issue of hand-flying proficiency among long-haul pilots. In this augmented crew environment the autopilot is off typically 3 minutes on an up to 12 hour flight. Pilots performing these rotations can have a hard time just staying current (i.e., an average of 1 takeoff and landing per month).

      I venture to say that even when hand flying, most Airbus pilots almost never turn off the flight director. The flight director does almost all the brain work in those cases. When I encourage my FOs to practice with the FD off, it is enlightening to see their reactions – as it is a different ball game. Following the FD bars telling you to pitch to 5° is not nearly the same as determining the 5° yourself—in fact, you probably don’t even take note of the commanded pitch attitude. I like to say that hand flying with the flight director on is like copying someone else’s homework. It’s your handwriting, but not your math.

      Now, complicate that with the fact that after the autopilot disengaged, the crew did not turn off the flight directors (part of the unreliable airspeed procedure). The flight directors turned on an off several times, but even worse than having no flight director, for much of the 3000 foot climb FO Bonin was faced with a flight director giving erroneous instructions. They were often giving guidance to continue the climb at the unsustainable rate of 1400 ft/min! For a guy of his experience, that I would venture to say always flies with the flight director on – and it is virtually always dependable, to ignore the flight director’s guidance would be very difficult. They are a very powerful visual cue, and there is much evidence in the report that he was trying to comply with them, all to the detriment of the airplane’s energy state.

      I must say, that I had higher hopes for Bonin’s basic airmanship and reaction to the prolonged stall warning as he was also a glider pilot—though we don’t know the extent of that experience. Gliders have no stall warnings at all—if they did they’d be going off most of the time! The reaction there is to simply push forward.

  58. susan says:

    unfortunately, Patrick, I think that just as in my profession (medicine) the actual skills of the trade (whether it be stick-and-rudder or physical diagnosis) have been subsumed to meeting deadlines, “throughput”, being “brand ambassadors”, “managers”, “customer service champions” and servicing the bottom line.

    I too read Langewische’s article carefully. I think the point he was making is that in today’s cockpit there are fewer and fewer who have HAD to rely on those basic flying skills, in large part BECAUSE of automation. and probably a smaller number who have that unique combination of skill and temperament that we hear about when an extraordinary disaster is averted (cite Sullenberger as sorta the modern-day archetype, but any number of that first wave post WWII flying aces will do). neither one of our jobs commands much respect anymore, partly because of the “anyone can fly” notion!

    • bobbi says:

      Fledgling “heavy” pilots need only master three books:

      1. “Airliner Piloting For Dummies”

      2. “Idiots Guide to Flying Jets, Helicopters and Transport aircraft.”

      3. “Learn To Fly Any Airliner in 21 Days.”

      Alternately, if that’s too much bother, just pop the three bucks for the “Cliff Notes” for each the aforementioned titles.

      Bada Bing, Bada Boom!

      You’re a transport pilot!

  59. Alan Greggs says:


  60. Don Beyer says:

    Invite Mr Langewiesche to sit with you in a simulator and teach him a few basics and then let him have at it. After all, ” just about anybody can fly these planes”. It’s inexcuseable that he did not consult with a pilot before writing such nonsense. This would seem to violate the most basic rule of his profession.

    • Rod says:

      Actually (and speaking of fact-checks), Langewiesche is a highly accomplished pilot in his own right and has written wonderfully on the subject.

      • Don Beyer says:

        Rod, thanks for the correction. Wonder why he would right such an article?

        • Rod says:

          Yes, he can’t have it both ways. His book “Fly-by-wire” is nevertheless well worth reading. As are “Aloft” and “The atomic bazaar”. He certainly is a first-rate writer.