Boeing push to make training profitable may have left 737 Max pilots unprepared
bloomberg.com
bloomberg.com
It's easy to sit in a conference room and say "gee, IT hasn't made any money, instead they've spent a lot of it! Let's downsize, outsource, and make our own divisions pay for IT, we'll see whether we're successful by measuring the IT department's profit next quarter." Inevitably, that department makes more money under that regime, but the company is worse at flying people from place to place.
If you're a hardware company, the software should be free. The support should be effective. You should be hosting training events at nice facilities with free pizza because people who know how to use your hardware will buy more of it. If I have to pay for your software, I'm going to be pressured to buy the Light version instead of Enterprise, and I'm going to pay for only one training class and not learn about your next iteration, and you will sell less hardware.
The only people they really care about and are accountable to are the shareholders and investors. And they want infinite profit at the sake of everything else. If you want to solve this problem then you need to remove them from the equation entirely because given enough time they will run a company into the ground and run away with the proceeds.
I think shareholders of Boeing would also expect stable growth through sound risk management
Nickel and diming on safety features and training is playing with fire.
Share holders of a big, established company want long-term and stable profit. Short-term profit at the cost of increased risk is not something that share holders of a company like Boeing are looking for.
On the other side, it happens all the time that a company has a single blowout quarter and the shares fall.
There’s an almost religious belief in whatever keeps the analysts happy this quarter is the right move. While this is done in the name of the shareholder, it happens to align with the compensation of the managers and directors making that decision.
I see Boeing as a defense/aerospace version of HP, GE, or IBM. In the case of Boeing, management went so far as to move a couple of thousand miles from most of the company. All of these firms chose to put themselves on the path to decline, and did so in a way obvious to anyone who looked closely.
Perfectly rational investors with infinite lifespans might do what you say, but in reality most people are fickle, short sighted and don't actually do much thinking.
As an example, when I worked for IBM, they furloughed the entire Systems technology group (essentially the folks working on things that aren't pure software or contracting) for a week.
This was done purely to reduce the payroll for the STG so that it's income would look better on the quarterly earnings.
Neither the inability to get anything done for 2 weeks (furlough was staggered so only approx half of folks were out each week) nor any of the follow on effects on project or work force was not enough to impede this type of roughshod pursuit of quarterly accounting numbers.
I mean, if you want "long term thinking", does that mean that WeWork is a better model than Boeing? Because people seem to hold it up as a failure of capitalism just as much.
A bond holder doesn't get paid their interest payment, it's a crisis. The government doesn't get the taxes due, it's a crisis. An employee doesn't get their paycheck, it's a crisis. Dividends are an "if we feel like it" thing. Many, many companies simply say "we have never paid any dividends and we may never pay any at all".
Stockholders by definition can only rely on the nebulous entire future of a company for stock value - either it will be sold some day for $X, or the dividends will total $X. That's a longer term perspective than anyone else has.
Not so fast - scroll through it, to 2019: https://en.wikipedia.org/wiki/Aeroflot_accidents_and_inciden... Be sure to expand sub-articles.
Instead, this is not really capitalism at work but internal forces inside a giant organization. The risk reward calculation for internal management does not track the company’s overall risks vs reward resulting in company wide losses.
It's supposed to be the responsibility of society to keep incentives in check such that this doesn't happen (via regulation or otherwise), but once profit-seekers leverage their profits to seize the reins of power you've got a big problem on your hands.
A) Not everything has to bring you a profit, jeez.
B) You know how you profit from training? By actually providing solid training to pilots who then fly your planes skillfully and don't crash and don't have any trouble landing. You not only get money from customers paying you for those pilots to fly them, you also save money by not getting sued into oblivion for your negligence.
"Let's make training profitable" + every other decision on the software side = 346 dead people. Those deaths weren't an accident, they were a result of Boeing's desire for perpetual growth at any cost, and to try and beat Airbus at their game. No one except the families of the victims will suffer from that. But go ahead and carry a bag of weed and be prepared to spend your life in the system.
Maybe it's time we rethink how limited liability works; and we get rid of this notion that corporations are beyond reproach (and they are, because the fines cost less than the profits).
One of the responses was to restrict capitalism (it became fettered vs. unfettered at the point). And it was a good one, one that has seemingly been forgotten or ignored.
Unfettered capitalism, through history, is a net loss for society. But the more we deregulate and seek profit at any cost, the more we re-align with that old view.
In this instance, Boeing seems to be ok with the death of ~350 people.
So it’s best not to use “capitalism” in the leftist sense and expect people to understand you. I also think there is a very serious and legitimate concern over the power of the state. The US state killed the leader of the Black Panthers, for example. We need regulatory control in the hands of the people. You may do well to study leftist critiques of the state. Mikhail Bakunun was active at the same time as Marx and was a critic of Marxism. Bakunin famously said in 1873:
When the people are being beaten with a stick, they are not much happier if it is called "the People's Stick"
For capitalism to work as intended, there has to be "perfect information", i.e., no lying and everyone knows what they're paying for. This is obviously a practical impossibility: companies and people lie, and most consumers buy things without fully researching or knowing what it is that they're getting.
One purpose of regulation is to improve the situation. If the FAA certifies a plane as safe, then consumers can rely on that certification and don't need to do their own safety research. They still don't have "perfect information", but it's a lot better than without it.
Good regulation is not "antithetical" to capitalism, it is often necessary for it.
If you assume that capitalism is intended to work not in the way sought by the people who drove the move to capitalism but instead according to idealized economic theories much younger than capitalism, sure.
I actually agree with you. Much of the foundation of capitalism was based on flawed assumptions ("perfect information" being a major one). Huge amounts of abuse were allowed to be perpetuated as a result, and it's fair to call into question the legitimacy of the entire enterprise.
That said, there are good ideas in capitalism, and we shouldn't throw out the baby with the bathwater.
Perfect information isn't part of the foundation of capitalism.
It's part of rational choice theory, which didn't exist until long ashtray capitalism. (There are arguments for capitalism rooted in rational choice theory, but those are retroactive justifications not the foundation of the system.)
> Huge amounts of abuse were allowed to be perpetuated as a resul
The abuse predates the theory “justifying” it, the theory isn't the cause of the abuse. Capitalism historically wasn't grounded in theory, it was grounded in the opportunism of the mercantile class under late feudalism leverage wealth, power, and the opportunities created by new technology to rearrange society so that the mercantile class (the subset that became the capitalist class) could displace the economic (and also political) power for the feudal nobility.
I claimed that regulation is antithetical to free exchange -- if you restrict me from trading something then I am not free to trade it, and if you don't restrict me then the thing is unregulated. I consider this statement a tautology. I don't know what you mean by "good regulation" and "capitalism", but I'll stand by my original phrasing.
Not quite. If you buy something without knowing what you're getting, that's not free trade. An airline selling a ticket on an unsafe plane but calling it safe is fraud, and fraud is not free trade.
A "perfect" regulation would only limit fraudulent trade, but allow all other types of trade.
However, because we don't live in a perfect world, regulation will inevitably limit some legitimate trade as well as fraudulent trade. A "good" regulation is one that limits enough fraudulent trade to more than compensate for the legitimate trade that it inhibits.
As a matter of language, I disagree. If I want to bet that the man offering me a lottery ticket is accurately conveying the odds, and you don't want to let me make that decision because you just overheard him discussing the fraud he was about to commit against me, physically stopping us from making that exchange by use of force is a restriction of my freedom to make that bad decision. We can agree or disagree about the desirability of that, but linguistically it seems pretty clear to me that this is a restriction of free exchange. Even when I supported such a use of government force I could recognise that.
If someone wants to sell your friend a lottery ticket, and you know with certainty that the seller has no intention of paying out a winner, this is not free exchange. It's just a lie and simple fraud. Furthermore, as a friend, you should probably at least try to intervene. The same is true of government (or any authority).
Making a fraudulent trade is not "free exchange", legally, morally, or linguistically.
In any situation where you are using violence to stop two people from exchanging physical objects, you're interfering with free exchange. If you actually disagree with this, I would like a term which refers to the concept I'm trying to discuss. I'd suggest that you're referring to something more specific like "well informed free trade" or "information symmetric free trade".
You seem be assume that "free exchange" simply means no one is physically stopping you from making a trade. That isn't what free exchange means. Free exchange means that two consenting parties trade known goods without undue influence. If someone puts a gun to your head to buy something, that isn't free exchange. If someone sells you a fake lottery ticket passing it off as real, that isn't free exchange. If someone sells you something they don't actually own, that isn't free exchange.
Meaning, when you get a parking ticket, that's state-sanctioned violence, because ultimately some non-consentual means could be applied to force you to turn over the money. Taxes are violence by the same argument, and, I presume, regulation of a lottery. (Someone might be sent to prison).
Again, if you can't agree with my use of words then please provide a term you find suitable for the concept I'm trying to convey. I provided two terms I find suitable for the concept you're trying to convey, but you've provided me none despite an explicit request. To me, when somebody refuses to find common language it seems like they're using language as a barrier rather than a bridge. To disagree on a definition can be understandable, but refusing to provide any alternative looks like rhetoric rather than a legitimate attempt at mutual understanding. What do you call unrestricted exchange of items between individuals regardless of differences in knowledge or legal ownership? I want a term that will clearly convey this concept to you, and others who are using language in a similar manner. [Edit: how would you feel about "unrestricted object swapping" or "unrestricted barter"?]
> If someone puts a gun to your head to buy something, that isn't free exchange.
I don't think I've said anything that would suggest this example is a free exchange, but the other two which don't involve coercion through threats of force I will stand behind (as free exchange, not necessarily as nice things to do to people).
And of course advertising, a nefarious tool to lead people to make decisions which otherwise they would have not.
the pressure of WW1 going on in Europe
Airplane manufacturing has lots of problems. But it’s a highly regulated industry. “Unchecked capitalism” is a bad description.
Many times when there is discussion about issues like the environment there is such a comment blaming capitalism. I can assure you from my own personal live experience and from having seen huge changes after reunification, in capitalism-free East Germany the environment was orders of magnitude worse off than in capitalist West Germany. Don't blame human behaviors on capitalism/socialism/whatever.
And of course, in neoliberal growth capitalism, the pursuit of perpetual growth will mean that some very stupid things are going to be done because they please the system: planned obsolescence, burning unsold clothes to keep the brand valuable, etc. They would mine the earth dry if it would make them a penny in their quarterly reports.
A few years ago I did some consulting work for British Airways. It was interesting in that as far as the planes themselves, and the operational risk of the planes was concerned, the company had their shit together. Decisions made from a safety first point of view. Engineers making the calls. Everything else no object.
But ALL the software around the planes, from the ops to ALL of the b2c was outsourced to hell. Cue a few months later their entire datacenter went down for 3 days, then a massive security breach.
Their idea of agile was to release stuff once every 3 months in a 2 week window.
The thing that I found odd was that they didn't really perceive software engineering as engineering. They had time, money and patience on the aero engineering, but software to them was just another cost center.
Oh boy, that's a good one. Like learning to write and learning to paint are easy. You can easily get SOME result...
By the way, I have automotive experience. They do, and have to, pay well.
My sister has been trying to get a ticket to visit me for Christmas. Couldn't do it because British Airways checkout is broken, tried in multiple browsers. Didn't work yesterday, didn't work today.
I wonder what the cost of annoyed customers who now fly Virgin and call center salaries is compared to proper engineering?
Someone once told me, it’s because generals and admirals can see hardware, and fell like they got value for their money. Software, not so much.
Perhaps it’s something innate in humans. Remember how Beats headphones used to add weights to feel more substantial?
He was right, we didn't ship anything physical.
We just lined up electrons. :)
I had a relatively expensive - but very lightweight - set of headphones a while ago and I swapped them for a cheaper but heavier pair because they just wouldn’t stay in place.
I would even say it is the opposite: The heavier the headset the tighter its grip has to be so that it stays in place during head movements. A tighter grip means more pressure around the ears, which for most people means less comfort.
I could see that while you concentrate on the feeling of headsets you may have that same bias as most people probably have that heavier equals better (quality, more solidly built), but I doubt that is true for the rest of your brain and your body.
I tried to find something, anything of substance on this concrete subject but this time my Google-foo failed. Links to research would be appreciated, if somebody else has better luck. A possible confounder is that heavier headsets - if it's not a cheat like in the discussed example - might indeed be better built, but I think we are talking about similar built quality here that only differ in weight, and objective criteria, not self-reported biased by already known to exist weightier-is-better value bias.
I only found this, which is really just a statement: "Relation between weight and comfort?" -- https://www.headphonezone.in/pages/headphone-weight
The weight of objects has interesting effects on our judgement: https://www.discovermagazine.com/mind/holding-heavy-objects-... -- but obviously that does not mean the body sees weight the same, physiologically.
It seems the padding matters a lot more than the weight, but lighter headphones need less padding, so for a given price point the lighter headphones are more likely to be comfortable.
This was not really true, the original tear down that found weights was of a counterfeit pair of Beats headphones.
https://www.gizmodo.com.au/2015/07/are-beats-headphones-real...
https://medium.com/@BenEinstein/how-it-s-made-series-beats-b...?
Beats didn’t do this. Cheap counterfeit headphone manufacturers did.
Can someone explain why this is actually bad?
(EDIT: I'm being serious, so please don't just down vote it without explaining. For airline industry with legacy systems, I don't necessarily think it's bad, but maybe I'm missing something.)
There's nothing inherently wrong with this, if 3 months are considered "frequent" in your industry (true for mission/safety critical systems, or systems that should not be upgraded frequently for some reason).
The idea is to deliver small, non-breaking increments to production frequently, rather than infrequent, big bang releases.
Hopefully the Boeing fiasco is causing other engineering lead companies to refocus their priorities.
I often wonder if the cavalier attitude towards treating software as an engineered product is because the software industry is relatively young. I imagine a lot of the cowboy attitudes we see in software were also prevalent in, say, pressure vessel design 130 years ago before adoption of strict design standards became common practice. I’ve heard it said that software is very much in the cave drawing days by comparison.
A competent exec doesn't structure a business so that "everything is a profit center", as that drives specific behaviors, and you have to be careful about the accountability you put in place and how you're measuring those departments. Competent management should understand that.
That said, there's an annoying fetish from human beings to obsess over what's easy to measure, which is not always what's important. Profit being one of those things...
Turning it into a profit center earns him respect and promotions.
Metrics have become the goto into having a business nowadays and we min/max everything, squeeze every little bit we can out of things. Its an open market and anyone is able to do however he pleases with his company, but I hate it.
The most recent example that comes to mind is Activision-Blizzard where they exceeded their revenue and profits for 2018 but they went on and downsized the company letting people go and cutting down the overall experience. I hate when this happens really, especially when its crucial to safety like airplanes. I think Boeing should have gotten more than a slap on the wrist for the max's that crashed.
The universities and pop-business books has made everyone so obsessed with the business processes themselves and their management, particularly the stuff that has commonalities across various niches: finance, management culture, 'innovation', etc.
But the all of the various unique fundamental value propositions are what really drives the companies, everything else is just helping deliver and execute that, but the core of the business often goes to the way-side in favour of obsessing over perfecting all of the the small processes in between. An ideal organization should put its primary business front and centre, above the various small processes which it is composed of mostly just for practical reasons of running a business.
It's sometimes difficult to distinguish between the two but I do very much think that it's possible to lose sight of that distinction in the day-to-day running of things, especially once the business gets really larger and running the business itself becomes the biggest time-sink for everyone involved.
https://en.wikipedia.org/wiki/Obliquity_(book)
And TEDxWarwick talk:
E.g. we cut 50% of the cost of training this year vs. we improved the training and pilots are more efficient.
The second is the continuing trend - in both private industry and government - to atomise organisations by outsourcing important work. The immediate, formal needs of the organisation will be met by the contractual requirements but informal communications and long term development of people and institutional knowledge aren't paid for and won't happen.
That reflects the management trend of seeing themselves as the customer and not the developer (of the company). What better way to reduce the number of opinions that must be considered than to turn your employees into service providers?
1. Drive MCAS with only one AOA sensor.
2. Don't tell MCAS to look for (or even think about) bad AOA data or AOA disagreements.
3. Equip airplane with two AOA sensors as usual, but make the AOA disagree warning light a "value added option" that customers have to pay extra for.
4. Don't actually bother to tell pilots that they don't have AOA disagree warning lights.
5. Don't bother to tell pilots that MCAS exists at all.
6. Don't test MCAS subsystem to see what it actually does with bad AOA data.
7. Give MCAS a ridiculous amount of control authority, operating cumulatively over repeated applications to exceed what the pilot can manually override.
8. Change how the trib stab switch has been wired since 1966 and not tell pilots.
How any engineering team let this through is unfathomable. Don't blame the MBAs or the unions. Didn't these guys get the lecture about the Hyatt?
0. Reposition the engine forward on the wing such that the airframe is inherently unstable and needs software for stabilization.
MCAS is nowhere near fast enough to compensate if the Max were actually unstable.
Sounds like unstable to me. "I am not fat, I am just insufficiently thin".
Picture a pendulum that is hinged on a greased axle vs one that is hinged on a high-friction axle.
Both will reach equilibrium eventually (they're stable) but one will do it faster/with fewer oscillations.
The 737 MAX is longitudinally stable, but it doesn't meet certain nuanced requirements of the Airworthiness regs.
> what happens is that the stick force lessens as angle of attack increases at high AoA and low airspeed, whereas in order to be compliant with certification it has to gradually increase with increasing AoA, just like a 737–100.
Nope, sorry, that's not what stability and instability mean.
Stability means it will return to equilibrium if disturbed, but the issue with the MAX is that it requires a positive force gradient per the Part 25 Airworthiness Regulations, which the MAX does not exhibit in a specific part of the envelope.
Stability is not a binary condition. The MAX is longitudinally stable in every part of the flight envelope.
"Not airworthy" does not equate to "unstable". Those two terms are orthogonal.
Like fixing a hardware problem using software. See this video about the plane's tendency to pitch up at low speed and full thrust. That's "insufficiently stable"! https://www.youtube.com/watch?v=6p5sboD2oO8
The same argument could be extended to the FAA oversight functions being “outsourced” to Boeing which may have had a direct impact on all those items you listed
They're all obviously coloured by the sources/reporter as obviously (in this case) if your pilot/simulator was impacted by cost cuts, then you're going to pin those dead passengers on your pilot/simulator changes. I'm fine with that. I'd do the same myself.
What scares me though, is that even now the story is still a pile of anecdotal tumble-weed.
What do we know? 1) Enough passengers died for the planes to be grounded. 2) These deaths happened due to: 2.1) Larger engines being bolted onto an old airframe 2.2) Plane not being sufficiently updated to compensate 3) FAA being a bit too cozey with Boeing in certification.
"Mistakes were made"
What I'm not hearing is Boeing going through the stages on this.
Where did they make the mistakes? What should they have done? What have they put in place/restored to ensure it never happens again? etc etc
Boeing needs to try to own what it'll take to get their planes flying again. And that requires them to engage with the bazillion people they've pissed off - to defend why this particular cost-cut wasn't why people died.
I think they will need to meet the EASA at the Canossa Castle so to speak for that plane to fly again. The FAA so completely screwed up the EASA won't trust a word from them for a very, very long while -- not only did they let Boeing effectively certify their own plane but when the fertilizer hit the circular cooling device and the EASA served them a face saving opportunity by not issuing an EU wide grounding order immediately but individual member countries, one by one, banned the plane from their airspace without grounding them. And the FAA didn't budge! Finally the EASA said, fine, the 737 MAX is grounded. And now all the trust and goodwill carefully built between the two authorities is all gone. Getting the MAX flying again in the USA is one thing, getting it through the EASA after this will be an entire different matter...
From Wikipedia:
"The Road to Canossa...refers to Holy Roman Emperor Henry IV's trek to Canossa Castle, Italy, where Pope Gregory VII was staying as the guest of Margravine Matilda of Tuscany, at the height of the investiture controversy in January 1077 to seek absolution of his excommunication.
According to contemporary sources, he was forced to humiliate himself on his knees waiting for three days and three nights before the entrance gate of the castle, while a blizzard raged. Indeed, the episode has been described as "one of the most dramatic moments of the Middle Age"
https://en.m.wikipedia.org/wiki/Road_to_Canossa
I also really enjoy Norman Cantor's description in "Civilization in the Middle Ages," where he takes a more cynical view of the actions involved.
As I understand it, airlines or the faa only permit pilots to be licensed for 2 airframes at a time. So a 747 pilot can also fly 737s but not other commercial airliners. That's a problem because if the new boeing isn't a 737, existing pilots can't fly it without giving up one of the other two. Airlines dont want that as it makes introducing the new plane a massive pain in the arse and it means either dumping old planes or running a complex mixed infrastructure.
So Boeing effectively had pilots fly a 737 simulator that would make the actual Max fly when/where/how they wanted.
Except the simulator had a bug where if a single sensor failed it crashed the plane.
So the trigger of the issue is a single point of failure in a sensor but the fundamental causes are needing to not change pilot certification and/or only being allowed 2 certifications at a time.
They needed a new regional airframe design a decade ago. Something to iterate on for the next 50 years. And that could be introduced while the 737 was phased out. A smaller version of the 787 would have a been a good foundation.
What exactly to do about that, I'm not sure...
The FAA does not limit how many type certificates that a pilot has. However, large airlines have policies that likely do limit their pilots to flying two airliner types.
Source: commercially-rated airplane pilot.
Dear bloomberg, ny times and other major news outfits.
Please sort out a decent micropayment solution that works on all news sites without hassle.
I'm willing to pay you a couple cents to read articles linked to by HN and other aggregators. However, I will never subscribe because I'm not interested in the other 99% of your content.
end offtopic rant