HN Theater @HNTheaterMonth

Yeah this is right. Vox has a video on this exactly https://www.youtube.com/watch?v=H2tuKiiznsY

https://www.youtube.com/watch?v=H2tuKiiznsY

So you want to use their historical record as some kind of justification for their dismal record with the 737 MAX, as if one is related to the other?

Cold comfort to the 300+ lives, and families destroyed.

Watch this video: https://www.youtube.com/watch?v=H2tuKiiznsY

This isn't just some "oops, these were badly trained pilots" – the whole premise of MCAS is seriously screwed up and smacks of non-engineers running the place.

⬐ privateSFacct

There record is not dismal with the 737MAX.

If you look at where fault lies part lies with boeing and part lies elsewhere in the safety chain. In contrast to many other areas - the evidence of ill intent is relatively weak here. Even without prison Boeing is facing major financial impacts as a result of this issue (as it should).

It's looking like evidence from the airlines that proper maintenance was done may have been faked.

https://www.bloomberg.com/news/articles/2019-10-15/pictures-...

We know lots of maintenance issues unaddressed and repeated warnings by the plane itself that there were sensor issues were not properly addressed.

We know response was not ideal on pilot side which overlaps with some training and other items around stab trim cutoff and/or automation dependency.

These factors are partly why there have been no accidents in 737MAX in US despite lots of flying.

Yes - boeing should design a totally safe plane. Part of that is going to be designing plans to accommodate a wider range of pilot skill (what they are calling "future pilot populations") and to better accommodate maintenance and ground handling training assumptions to allow for greater risk of problems there. This is already being implemented.

Ironically, one element may be to REDUCE the reliance on pilots as a key flight safety control and then increase automation and redundancy in the automation.

⬐ thatswrong0

> the evidence of ill intent is relatively weak here

Most people are referring to negligence, not ill intent. "Let's make more money and rush this thing out! (Even though this could kill people)", not "Let's deliberately design a system to kill people".

Regardless, one sensor failing should not induce a plane to fly itself into the ground, despite the best efforts of the pilots to recover (and despite the fact they were not necessarily the best trained pilots). 2 / ~400 complete hull losses of brand new jets is completely unacceptable. That is dismal. Don't try to downplay it.

⬐ privateSFacct

These were not the "best efforts" to recover - stab trim cutout is a memory item.

Absolutely - Boeing is going to be designing much safer planes in future to accommodate different pilot populations. That is clear and necessary. This will make us all safer. Boeing obviously screwed up with a primarily US based mentality.

This system was PARTICULARLY fragile in the face of poor maintenance and reliance on automation - which boeing was unreasonably dependent given a US centric view.

Look into history of comets and concorde if you want to look at hull loss rates (note - both stopped flying forever when safety issues became clear). I predict almost no chance that the max will be taken out of flying forever.

⬐ bronson

By US-based mentality, you mean one where it's OK to design an automation system that takes input from a single sensor? And this is somehow OK for American pilots, but not for foreigners?

I think evidence will show you're wrong. The MCAS is a terrible design no matter where you intend to fly it.

⬐ privateSFacct

Yes - automation in the US context has historically evolved with pilots pilots considered a key and reliable redundancy. So yes - the automation is not full automation / full redundancy. The idea has (historically) been then that the pilots bring overall safety to required standard and the pilot is the redundancy. The evidence in the US has showed that this approach works from everything from stab trim cutout issues to dual engine failures.

In the US, efforts in last years have focused on the entire chain of safety (actually away from plane design) and the US has hit some of the BEST periods of safety in aviation. This is everything from duty / rest periods, maintenance, minimum training requirements etc.

"And this is somehow OK for American pilots, but not for foreigners?"

Let's be VERY clear here:

The requirement to fly Part 121 in US even as just a first officer requires an ATP and 1,500 hours. Most folks have a a significant amount of other flying prior to that, the US has a much bigger GA aviation scene, private flight instruction, glider, private jet / charter markets and military flying. The US has something like 12,000 pilots flying for the air force alone. The first officer on your long flight is very very likely to have a TON of experience in all sorts of situations, and particularly with manual flying.

Overseas - you can be a graduate with an academic / training background only and 250 hrs of "flight time" more broadly defined and be at the controls. Ethiopian airline pilots were 25 and 29 years old.

In the US, major carriers even more picky.

Southwest for example:

Flight Experience: 2,500 hours total or 1,500 hours Turbine total. Additionally, a minimum of 1,000 hours in Turbine aircraft as the Pilot in Command* is preferred. Southwest considers only Pilot time in fixed-wing aircraft. This specifically excludes simulator, WSO, RIO, FE, NAV, EWO, etc. "Other Time" will not be considered.

So yes, if boeing is going to sell into markets with folks getting into cockpit as co-pilots with sometimes as low as 50 hrs (1 week) of experience (in cadet / training systems) then 100% they will need to redesign and rethink things for this future pilot population - and yes, that means the systems will have to have MUCH more natural redundancy built in.

Ignoring this difference in background, opportunity and skill will lead to lots more deaths - all preventable if the thinking changes and a better understanding of future pilot populations is developed, and that includes accounting for differences in training and experience.

⬐ bronson

You hype up reliance on US pilot training, but ignore Boeing's intent to NOT train pilots on MCAS (and especially its differences from existing behavior). Plenty of US pilots have gone on record saying that their US-based MAX-8 training left them unprepared to fly the MAX-8 as it was originally delivered.

In fact, even the Southwest pilots you mention disagree with you: https://www.seattletimes.com/business/boeing-aerospace/in-sc...

You're still trying really hard to blame the pilots and ignore critical and unarguable airframe deficiencies. It hasn't worked yet but maybe if you keep writing lots and lots of words you'll get somewhere.

⬐ privateSFacct

And despite no training we have what may have been some amazingly good MCAS recoveries within the US during takeoff into things like turbulence triggered MCAS nosedowns - pilots thought their own fault / couldn't understand what was going on but the limited reports in terms of response are great for totally untrained situations. I'm sure pilots were pissed, they should be!

I think you are failing to understand how the chain of safety works. Ideally you have a great design. If you don't then ideally the flaws / weaknesses don't get hit because you have things like good maintenance. If you don't have that then maybe next hope is the pilots. Worst case you then you require instrumentation to help reduce risk of the next tragedy. Etc.

This system has resulted in an incredibly safe mode of transport despite many challenges / crazy tolerances etc.

⬐ bronson

"Amazingly good MCAS recoveries" are a crew success, yes, but they are even more a sign of aircraft and regulatory failure. You aren't acknolwedging that the MAX-8 has serious issues no matter where the pilot learned to fly.

American pilots are top notch. However, the pilots themselves don't want to be put in the position of saving the aircraft, and a big part of the FAA's charter is to ensure that they don't have to.

Does your chain of safety include feeding a twitchy automated system from a single sensor?

⬐ thatswrong0

> stab trim cutout is a memory item

1. MCAS does not behave exactly like a normal runaway trim.

2. On the Ethiopian crash, they _did_ initially follow the stab trim cutout procedure, but they didn't manage their speed correctly and thus were unable to control the plane with manual trim.

IMO point 2 is very important - pilots will not always be perfect. It's easy to armchair pilot from your couch and say that "Oh, just manage your speed correctly when your plane is trying to pitch itself into the ground just after takeoff and follow the stab trim cutout procedure. Very easy to deal with MCAS." But reality is not that easy.

> Look into history of comets and concorde if you want to look at hull loss rates

Compared to every other new plane released in the past 20 years, the 737 MAX has a terrible hull loss ratio. I don't really care about the Comet or the Concorde since it's 2019.

⬐ privateSFacct

Agreed - pilots will not always be perfect, and if the safety chain -> design of systems, maintenance, training, experience is weaker than this will matter more.

Pilot error still #1 cause of fatalities including in the US. In this case bad maintenance also looks to be an issue and a brittle design. You have a master caution on 4 seconds into flight - you've got a terrible maintenance log book etc. I actually thought initial stab trim call wasn't unreasonably delayed - though they did let it run 4.6 to 2.1 over 10 seconds or so, which is a HUGE trim change.

And stab trim cutout wasn't followed fully or MCAS wouldn't have been able to put them into the ground at the end, though they may have already in a bad spot so thinking not unreasonable, except of course you might sit on nose up trim for a bit once reversing cutout.

The move will be I think to more automation and reducing ability to dispatch with things like a flight computer out of service (histrionically pilot skill would cover this now).

I also think they may need to move to authenticated maintenance / self test / parts life / history items. Ie, prevent dispatch on sloppy fixes / sloppy test.

⬐ bronson

> Their record is not dismal with the 737MAX.

By what metric?

By every metric I can find (passenger deaths per trip/flight leg/flight mile, hull losses per delivery/year, etc) the MAX is an outlier.

In fact, I can't find any other airframe in the last 50 years that even comes close (including the DC-9 and TU-154). What other airplane has killed 300+ passengers in its first 4 years of operation?

So, what metric have you selected to show that the MAX isn't dismal?

(and please, no lie-with-statistics stuff like hiding the MAX in 737NG data, or claiming that crashes by foreign pilots or on foreign soil don't count)

⬐ LifeLiverTransp

Sometimes a place can fall apart from the inside, knowledge rot eating away at a engineering culture, like termites at a house. In the end, its all still there- but not reliable.

MCAS is ducttape. Ducttape applied, because the whole culture is by now incentivize screwed, pennywise optimizing instead of pound problem solving. A "carry on" is actually the hostile action in this situation. Sidenote: I like that you can not lie in modern society, because all data has copies and copies of copies.

⬐ Gpetrium

What op seems to imply is that airplane incidents rates actually have a stronger track record as opposed to the opioid epidemic, road fatality, etc. And that if society believes that this individual/group should go to prison based on this mistake and in hindsight, without account for all the other factors that were a part of the decision making process, then society should start reconsidering the way it treats opioid, road fatality, etc.

The mindset of "quick to imprison" can also run the risk of creating a society that is overly averse to risk taking, which can hinder technology and scientific advances. For example, it may take 10 times as long to get a new, more advanced traffic light implemented in your city because now everyone wants to make sure no stone was left unturned, otherwise someone will get into an accident and a staff/group will be imprisoned. Or a new software is implemented but 3 months later it is found that failure under very specific scenarios has caused over 50 deaths. There are millions of potential scenarios that may fall under similar conditions as exemplified above.

Please note that this writing is not advocating for or against either views, it is simply shedding light on risks that should be considered.

⬐ luckylion

Has anybody argued that nobody else should be punished, ever? Otherwise "but there's drug addiction and car accidents! We shouldn't pass judgement on Boeing until everything else is perfect" is a really strange argument.

⬐ privateSFacct

The really strange argument is saying it is totally OK for people to willfully engage in activities that result in significant and ongoing fatality rates within the bounds of US law without punishment (slaps on the wrist for everything from willful pollution to opiods where the investigators were waved off) and then demand jail time for folks who have no fatalities within US law AND have an incredible safety record in their field, a record FAR FAR better then lots of other areas (drug distribution, medical malpractice, enviro health and safety etc).

If you wanted to reduce auto accident rates, opiod deaths etc you'd put these folks in charge, not put the law enforcement lobby in charge (yes, they will arrest lots of low level offenders but will not systematically address the issues and do not chase the folks at the top).

⬐ privateSFacct

I'm making the point that for for US flying (with US levels of maintenance / pilot training etc) boeing and even the 737MAX has a safety record that is incredible.

Not only that, it beats almost all other regulated modes of transit and even other regulated hazards (OSHA controlled worksites etc).

As always, it could be better - but it's actually amazingly good already - these planes are incredible safe in a challenging environment (miles, landing cycles, tolerances etc).

The demand for prison time here, when we have so many many areas where prison time can be MUCH more closely and immediately linked to bad actions (and goes unpunished) is misguided.

⬐ situationista

If it were safe to fly the Max in the US, and the problem were the pilots, it would be in the air now. It isn't, as the problem is clearly with the plane. It was likely just a matter of time before another Max fell out of the sky somewhere in the US, Europe or elsewhere.

⬐ privateSFacct

The problems were a chain -> starting with terrible maintenance not pilots.

And part 121 planes in the US do not fly multiple legs with the problems described on these planes.

I know it's partly a labor issue, but mechanics at united will take a plane out of service for practically any reason.

⬐ WalterBright

This is already the case with the FDA 1962 "safe effective" regulations. There is little downside to the FDA wanting to take no risks whatsoever in approving a drug, but a huge downside to the regulators if they approve a drug that turns out to have a fault in it.

The result is that developing new drugs got enormously more expensive, far fewer new drugs get developed, long delays in effective treatments getting approved, diseases that don't affect large numbers of people don't get cures developed, etc.

The net result was a negative for patients.

This was all discussed in "Regulation of Pharmaceutical Innovation" by Sam Peltzman.

There have been a lot of deadly aviation crashes due to mistakes, false assumptions, oversights, incompetence, human failings, etc. But somehow we've wound up with incredibly safe airline travel. Millions of flights with no incident. Do we really want to start jailing people now? What improvements will we forsake if we give airframe makers powerful incentives to hide mistakes? or simply avoid making improvements to safety, because who wants to risk jail for making a mistake?

VOX has a good explainer video on what caused the 737 Max to crash twice for anyone who doesn't know the details. Essentially, Boeing installed sensors and software to automatically push the plane nose down if the plane angle was too steep as it ascended. This software/sensor combination is implicated as the cause of the recent crashes in Indonesia and Ethiopia.

The real reason Boeing's new plane crashed twice:

https://www.youtube.com/watch?v=H2tuKiiznsY

⬐ ethbro

At this point, I think everyone on HN who doesn't specifically block Boeing articles is aware of the reason.

Meanwhile he claims 737 will be their safest plane ever: https://www.youtube.com/watch?v=DE4tX0eIDEQ

Edit: unsure if this video is correct, but the MCAS system is classical "solution" to how in software a junior programmer would patch their software by adding more complexity to the code without realizing what that changes would further affect. It felt like they gave up on some solid solution and just decided to added a sensor who would overwrite pilot manuver.

A similar situation would be downgrading a size of tires on a new model of a car (to be competitve with other auto makers that lower the price) and simply adding software blockage that you cannot turn wheels as violent as you could before because on thin thread that would lead to a car tipover.

The whole video - if correct, paints a grim picture of profits above PAXes...

https://www.youtube.com/watch?v=H2tuKiiznsY