Comment by wolvoleo
16 hours ago
Hmm I don't think it's as black and white as just blaming airbus. The pilots literally flew a perfectly flying plane straight into the ocean. And they had plenty of time to understand what was going on. But they didn't. They didn't willingly do it and the system misguided them but that wasn't the only factor.
I agree airbus shares the blame but it's not the only one. The pilots should have realised the situation they were in, their training should have been better, there were a lot of factors.
Admiral cloudberg has a good deep dive on it. https://admiralcloudberg.medium.com/the-long-way-down-the-cr...
There were other near accidents before due to the exact same problem, the problem was well understood, and the changes needed to solve it was known.
Air France didn't implement them and Airbus didn't require them because of money. They thought the chance of it causing a real accident was low and decided to risk it. Despite there being known near accidents already.
And yes, "[the pilots] training should have been better" is part of the things that put both companies at fault. It's not the pilots fault that their training didn't cover it.
> Airbus didn't require them because of money
I am pretty confident that aircraft manufacturers themselves cannot require these things, only regulators can. The FAA in particular used to lean heavily on budget constraints for airlines (who would also push back against expensive upgrades); but I am sure the same applies to EASA and other regulators as well.
The manufacturers literally write the manual. The regulators only approve or reject it. And yes, EASA approved it too.
They should be able to recall a plane for a safety flaw. In which case they have to pay for the upgrade themselves.
If the airline doesn't comply afterward, it would be on them.
But they didn't issue a recall, so they wouldn't have to pay for the fix, an over 200 people paid the price instead.
At least, that's how I read the blame distribution.
7 replies →
That's right, Airbus is responsible for the faulty equipment onboard, not pilot training. Air France is responsible for its pilots' operational training and recurrent training.
2 replies →
Separating "regulators" and "manufacturers" in such distinct categories is overly simplistic, I'm afraid. As we saw with the whole Boeing debacle, the manufacturers are the experts on what they build, and we expect them to give clear, levelheaded, and honest guidance to operators and regulators. That also means they must have some responsibility for the outcomes of that guidance.
Having a separate regulator, which does no building themselves, somehow maintain a separate team of independent experts is a fools errand. We should of course have independent evaluators, but the people building the thing are the experts on the thing.
1 reply →
>There were other near accidents before due to the exact same problem, the problem was well understood, and the changes needed to solve it was known.
Could you be more specific here? The article doesn't even say which problem Airbus are considered to be criminally liable for.
The plane was prone to lose its airspeed sensors on some weather conditions.
As much as I understand the culture of blameless post-mortems and the fact that people in that cockpit don't get the benefit of hindsight, maybe those other companies didn't have an accident because they followed procedure (which was a simple one)
Yes there were UX factors. Yes training could be better. Yes distractions happen
But if I'm going to blame the companies I'm going to blame them on putting someone inexperienced and probably who did not have the right mindset in navigating the profession. And meanwhile companies waste time in making automations on top of manual processes that make things even more complicated
Such an incredible write up, the piece about the importance of flying less technological planes to get a "sense" of what flying really is hits like a brick, specially in the world of LLMs producing code.
How do you get this "sense" of writing code and building systems by yourself if all you do is instruct some agent to do it? Are we all going to be like Bonin in the future where we just don't understand anything outside of the agent box?
This is both terrifying and sad.
I'm a software engineer and recently got my pilot's license, and the training for the pilot's license increased my (already-high) respect for the aviation profession. All pilots learn to fly basic airplanes and have to do everything by hand (often on paper, but an iPad is allowed) to show they know the basics. The result is that by the time you work up to more advanced planes you have climbed the ladder of abstraction and know what underpins the automation.
The other piece of the picture is that pilots acknowledge that their skills are perishable, and they have to commit to ongoing training. This would be analogous to writing code by hand and getting a licensed engineer to sign off on your currency periodically even if you use LLMs for work.
But I mean flying a cessna vrs something that has fly-by-wire like Airbus jets, its not really about understanding abstractions or anything, since the plane is basically a fundamentally different machine no? Basic principles of gravity and physic apply sure, but the flying experience is 100% different and not like a levelling up thing right? Like i would not trust someone with a Cessna pilot license to fly the airbus i am on.
7 replies →
But it wasn't at all just about Bonin: Robert and Bonin repeatedly kept trying to override each other; Robert was giving Bonin some information with which he could have figured out he was stalling (although Robert was also trying to climb); and Dubois had gone to deal with his sleep deprivation without designating either of them as PIC, and when Dubois finally returned (2:11:40, sounds still sleep-deprived based on his confusion) he didn't recognize the obvious stall or take control until they had lost almost all of their altitude.
It makes Air France look worse that all three pilots didn't react properly (or weren't trained or experienced), than just faulting Bonin. And at least two of them were sleep-deprived. But there were multiple systemic failures, not just the pilots.
Novella "Profession" by Isaac Asimov.
"Profession" is often cited with regard to LLMs, but honestly, in reminded more of (and scared by) "The Feeling of Power".
The irony of not understanding almost 100% of the code on modern airplanes is actually done by instructing a program to actually generate the code. It is neither terrifying nor sad. You expect humans to write millions of lines of code? At that scale, procedureally generating code is much safer and smarter.
Those millions of lines of code can often be reduced 10x or 100x with just a bit of common sense, and with that also reducing the potential bug count by 10x to 100x.
Also unlike LLMs, traditional code generation techniques are deterministic.
I'm not flying anymore if that's the case.
Actually there are more planes flying today than ever and the number of accidents is very very low, thanks to technological planes and protocols that lean from mistakes.
So low in fact that the majority of the recent "accidents" look like suicides from the pilots. The pilots know exactly what they are doing when crashing the planes.
Boooo!
> Hmm I don't think it's as black and white as just blaming airbus.
Then it’s a good thing they didn’t. Both Airbus and Air France were found guilty, and poor pilot training was specifically called out as a reason why Air France was considered guilty. It’s in the article.
I mean, one doesn't even need to read the article to learn that the blame went to both companies - it's even in the title here on HN.
Is this the crash where the pilot failed to recognize the airspeed sensors had frozen up and he stalled the plane? I could see how this was an Air France fault since the pilot was not properly trained or experienced to fly this plane in these conditions. Not sure why Airbus is responsible.
it's the crash where pushing nose of the plane down (correct enough-altitude stall response) caused alarms to activate, while pulling nose up caused alarms to silence
no wonder airbus was found guilty
Airbus kind of embodies the "trust the computer" mentality; and if you're going to do that the computer damn hell better be right all the time - it must not have "backwards" failure modes.
Boeing, in similar situations "in the past" would just sound a "computer is giving the fuck up, fly this pig dog" bell and leave it to the pilots to figure it out.
7 replies →
The behaviour you describe above only occurred after the pilot flying stalled the plane. There was a procedure for unreliable airspeed indication. Had the pilot flying performed it, the situation would have been resolved without incident.
AF could perhaps be held liable for insufficient training on high-altitude stalls or recognising and responding to reversions to alternate law. But it's hard to see how Airbus can be responsible for a pilot ignoring the most basic first response.
2 replies →
Thank you, this accident reminds me a bit of the McDonald's coffee lawsuit, where the popular narrative of "be less of a dummy" is not really fair
Edit -- to wit: https://news.ycombinator.com/item?id=48253931
While true, pilots aren’t trained to just “respond to the alarm” they are trained to fly the plane.
Once there were multiple alarms that made no sense at all (petty early in the event), the pilots should have ignored them as per the checklist.
But the most damning thing is the one pilot pulling the stick back and holding it back for almost the entire event. There aren’t any flying conditions where that’s an appropriate input. Not to mention being told to give up control and ignoring that request.
I agree Airbus has some blame in terms of the computer system not adequately communicating when it drops out of normal mode.
2 replies →
It reads exactly like "Ironies of Automation" by Bainbridge would predict.
Yes, an autonomous plane would have worked so much better. Can’t wait for AI to replace stupid apes.
A crash instigated by failure in software automation inputs would have been better handled by full AI software automation?
I actually think that is likely. Humans in these conditions have to make decisions under immense stress. Machines don’t, they just need to be able to understand that sensors may fail and are not completely reliable all the time. Though they would need lots of different input , just like humans, to be able to call out which part of its input is flawed.
2 replies →