By design, a fighter pilot does a heck of a lot more than a airline pilot. When a pilot does something wrong, we rarely hear about it.
Comparing them in this manner is just disenginuous.
True, a 747 pilot will have in the vicinity of 400 souls on board, while a fighter jet pilot will have 1 or 2.
That said, you're breaking the site guidelines: https://news.ycombinator.com/newsguidelines.html
>>usually a little higher than the “middle school”
You too.
> I converse here on HN because the discourse is usually a little higher than the “middle school” level that you find elsewhere on the Internet.
Rather than your edit:
> usually a little higher than the “middle school”
Edits like these don't inspire good faith discussion. Frankly, I think this thread has thoroughly outlived its useful life.
If you think the Su-25 is a bomber then the A-10 is one too ;)
Both are subsonic attack aircraft's, completely different missions.
Those are Bombers:
https://en.wikipedia.org/wiki/Strategic_bomber
And since you mentioned Ukraine:
https://en.wikipedia.org/wiki/Tupolev_Tu-160
https://en.wikipedia.org/wiki/Tupolev_Tu-160#Former_operator...
Pretty interesting backstory.
Automating flight has quite a bit of low hanging fruits and Airbus has become the first harvester out there. The rest opposed at first, but the only logical path forward is automating people out of the systems. Yes, we had the 737MAX thingy but even then arguments for re-humanisation of aircraft controls didn't catch much traction.
Is that sure?
Autonomous cars are hard because of “surprise”. Other drivers, kids, bicycles, randomness intersecting the street that has to be dealt with.
By and large, aircraft have the way cleared for them by ATC, and objects that may intersect with them tend to be on a path that is neither “came out of nowhere” nor erratic (UFOs aside!). Weather is a challenge, but autopilot deals with weather already. Arguably there’s much less surprise for a sky bus than a road bus, and arguably the fact it’s a big airliner instead of a crossover SUV doesn’t make it particularly harder for a machine.
TL;DR: Planes seem “nearer” than cars and some cars pretend to be automatic already.
I'd imagine there would be a many-decade long period of overlap where pilots would observe the plane flying before the plane can fly solo.
If MCAS _actually_ worked, you and I wouldn't even know about it. But as tribute to how poorly it was designed and delivered, we know about MCAS because it killed over 300 people.
Just an absurd point you're making.
Airbus are an engineering firm, so they are doing clever engineering things. Boeing was an engineering firm that has been run by bean counters the past few decades.
It wasn't "opposition to automation" it was opposition to innovation for the sake of short term profit.
Boeing's executives should all be hanged :) hundreds of people dead for no reason whatsoever.
Boeing was going to start a clean sheet small aircraft. And then American Airlines announced they were going to buy 100 of a new 737 variant.
The MCAS system was built to work on a very buggy angle of attack sensor. They are notoriously unreliable. The max kept one on each side of the plane (the airbus equivalent has 3 on each side). Further compounded by the fact that the (this was true at the time and theyve joined systems now) pilots computer only used the left side of the plane instruments and the copilots computer only used the right side.
So no. Boeing fucked up in the first place by trying to stealth fix a hardware problem (stalling at a new lower angle of attack in the "same model airplane") with software. And then lying through their teeth about the impact of this.
The equivalent military 737max with MCAS simply allowed the pilot to more forcefully adjust the flight control stick to turn it off. So they knew better.
Companies attempting to skirt legislation in the face of better profits is always going to cause issues here.
The entire max disaster was an attempt by management at Boeing to get pilots to be able to fly all 737s with basically no retraining.
> only used the left side of the plane instruments and the copilots computer only used the right side
These aren't really "pilot" and "copilot" computers, but two redundant computer systems (each with redundant processors) that take turns running every other flight.
> trying to stealth fix a hardware problem (stalling at a new lower angle of attack in the "same model airplane") with software
Actually, trying to stealth fix a human-related hardware problem (stick feel and overpitching tendency, not stall) with a crappy software approach.
> And then lying through their teeth about the impact of this.
A big factor here is that runaway trim was considered a serious problem but also one that flight crews regularly train for and are tested on and are expected to handle. But MCAS gave this exam to pilots far more frequently than runaway trim had happened before and we found that crews often fail the test.
The system design didn't try very hard to prevent this occurrence, because it was thought that flight crews could handle the problem when it occurred.
> The equivalent military 737max with MCAS simply allowed the pilot to more forcefully adjust the flight control stick to turn it off. So they knew better.
It's not really the same system, and it's not a "military 737max"-- I think you're referring to the KC-46, which is built upon the 767, and has a pitch feel control system also called "MCAS" that solves a related problem but overall operates very differently.
> The entire max disaster was an attempt by management at Boeing to get pilots to be able to fly all 737s with basically no retraining.
This worsened things, but the main issue is that there is a certification requirement -- https://www.law.cornell.edu/cfr/text/14/25.173 -- for pilot pitch feel on climb which is hard to meet with larger engines under the wing.
I'm not saying it's impossible that if they'd ignored the checklists and immediately made the perfect decision they could have turned back and landed the plane - just that it's far from certain, and if pilots get in the habit of doing this, it will harm more than it helps.
Computers make it technologically feasible to simplify things, but also to make them more complicated. Big organizations always inject as much complexity as possible into engineering projects.
They're not. Older planes required more personnel (flight engineers, navigators, etc.). Newer planes are vastly easier to operate with just a captain and first officer.
https://www.faa.gov/aircraft/air_cert/design_approvals/air_s...
Here’s the funny part at the end:
> The manufacturer admitted that desktop products in general are not good candidates for avionics, and the COTS components that are most likely to be effective are industrial-grade components with a wide distribution base such as medical-based products.
What’s ironic—I think he thought I’d be impressed…and that I would be more likely to want to fly in it. However at the time I was working as a Network Engineer (for him) and I while I was happy to entrust my livelihood to Microsoft, I would definitely not trust my life to Microsoft
OMG.
I thought for things where human lives are acutely at stake they audit every single line of code.
I mean, Windows has a known history of crashes and freezes and things like that.
I'd trust some kind of unix (BSD) a lot more than NT.
Tangential: For many non-consumer applications, I see plenty of situations where you'd think a Unix variant (or even bare metal electronics) would be an obvious choice, but no it's usually a software running on top of Windows.
Like, my local tyre shop has their wheel alignment software running on Windows XP. Before seeing that, I'd have thought it'd be some contraption with sensors attached to 7-segment LED display or something, which would also theoretically be able to run forever.
Or…asking me to troubleshoot the damn thing since I had a Windows NT MCP at that time.
(similar to "is there a doctor on board?")
A Cessna 152 (a single-engine, two-seat, low-performance aircraft popular for flight training) isn't very complex to operate. It can't fly at 30,000 feet, cross continents in a single flight, fly in icing conditions, land in zero visibility, safely take off or land in a crosswind that could be described as more than a stiff breeze, or carry much more than two people and a backpack. Airliners do all of those things. Military aircraft have similar performance demands, and people shooting at them.
Automation can probably simplify everyday operation more than it has so far, but virtually every system that can be put on an aircraft has failed in flight at some point so pilots have to be able to quickly identify malfunctions and manually override a half-working system. For a recent example of what happens when that isn't the case, see the 737 Max.
Perhaps it's inertia of regulations and training. Radically automating airplanes involves training loads of people. There's also certification, which is different for military aircraft than for civilian aircraft.