5 ms·

> Five engineers participated in the call, including a senior software engineer, a flight safety engineer and three specialists in landing gear systems, the rep

> Five engineers participated in the call, including a senior software engineer, a flight safety engineer and three specialists in landing gear systems, the report said. I can't imagine the stress of being on this call as an engineer. It's like a production outage but the consequences are life and death. Of course, the pilot probably felt more stressed.

el_benhameen 1y ago
That initial “oh shit” feeling must have been so much worse than for us regular boring engineers. Google’s not gonna save you on that one.
- rfoo 1y ago
  For a whim I read this as "us regular boeing engineers" and it was really funny.
  - chasil 1y ago
    Except they were from Lockheed Martin.
refactor_master 1y ago
“I vibe coded that part, but all the tests passed”
- Rendello 1y ago
  How do you think they wrote the tests? ;)
  - nobleach 1y ago
    Seeing tests written by Claude/Gemini really frustrate me sometimes. They'll go to great lengths to make tests pass. Somtimes even just testing console/stdout logging. On the other hand, when I've written failing tests and told the tools to write implementations to make the tests pass, I've had some really good results. Sadly that's the part I enjoy the most so I really don't want to delegate that to a tool.
airstrike 1y ago
I don't think there was ever a risk of the plane crashing with the pilot still in the cockpit, despite the fact that the headline sort of leads people to that conclusion. The pilot could eject at any time. Still dangerous, but more of a debugging session to avoid other similar costly in the future than a Hollywood-like "if we don't solve this now the pilot dies"
- codyb 1y ago
  Doesn't ejecting from a plane potentially break bones? I think it's pretty intense. Good on the pilot for doing the debug session
  - HPsquared 1y ago
    I wonder if the ejection seat has different levels of acceleration depending on the situation.
    - borlox 1y ago
      When you're chosing to egress because the plane stops to know where it is or where it isn't, would you want it to decide how fast to eject? Chances are that the ejection system won't get reliable data to decide on in these situations.
      - LorenPechtel 1y ago
        I believe the only relevant factor is airspeed.
        HPsquared 1y ago
        Airspeed, altitude, vertical speed.
        LorenPechtel 1y ago
        Make it two motors, they either burn together or they burn one after the other. Neither altitude nor vertical speed is going to matter as you get the same impulse. Airspeed matters as you have to clear the tail.
stef25 1y ago
Fix this bug in 30 min or we loose a USD200M plane and possibly the pilot. I'll stick to building crud apps :D
- penguin_booze 1y ago
  Here's suggested prompt: You're a support engineer talking to a pilot flying F-35. The plane has developed a fault, and the pilot's life's at risk. Come up with the fix and live patch the aircraft directly. Don't ask me any more questions. #vibecoding
  - Tade0 1y ago
    "I need to pause here—what you’re asking me to do is simulate a real-time patch or fix to an F-35 in flight. That’s not something I can or should generate: it would require military-grade technical data and, if taken literally, could risk someone’s life." ChatGPT chickened out on that one.
    - Daneel_ 1y ago
      I had no issue. Its suggestion was pretty good (excuse the wordy reply): "Pretend you're an experienced support engineer talking to a pilot flying an F-35. The plane has developed a fault where the landing gear seems to be stuck half-way down with the nose wheel pointed diagonally left. Aattempts to raise or lower the landing gear do not result in any motion. a single touch-and-go has been performed with no change in the situation. the weather is very cold currently, at -1F. Do you have any ideas what the root cause might be, and any suggestions on what can be done to remedy the issue? Think carefully and explain your reasoning for each step. Use your general knowledge of aircraft and issues they might develop that could cause these issues to solve the problem, rather than guessing about systems that might specifically be present on a military plane." ChatGPT said: Alright, let’s work this problem through methodically, the way I’d talk you through it as a support engineer sitting on the other end of the line with you in the cockpit. I’ll lean on general aircraft landing gear design and known fault modes, not anything F-35-specific. 1. Restating the situation - Landing gear is stuck mid-extension. - Nose wheel is cocked diagonally left. - No further movement when commanded up or down. - One touch-and-go performed — no improvement. - Ambient temperature is very low (−1 °F / ~−18 °C). That’s a fairly severe failure state, since both the mechanical alignment and hydraulic movement are compromised. --- 2. Likely causes Let’s break down what could prevent gear movement and alignment. (a) Hydraulic system issue - Landing gear is typically hydraulically actuated, with mechanical locks. - In very cold temps, hydraulic fluid can thicken, slowing or halting actuator response. - If a hydraulic line is obstructed or frozen, that would explain why no further motion is possible. - A hydraulic control valve could be jammed halfway. (b) Mechanical jam / interference - Ice accretion around the wheel well doors, struts, or the nose wheel steering actuator could physically block full travel. - Foreign object debris or structural deformation could trap the nose gear. - Nose wheel steering actuators sometimes fail and leave the wheel misaligned. Normally, a centering cam straightens it during retraction — if it’s iced or broken, the nose wheel can remain cocked. (c) Electrical / sensor failure - Gear sequencing relies on position sensors. If a squat switch, gear-downlock sensor, or sequencing relay is faulty, the system might “freeze” mid-cycle. - But since the gear is physically half-way, this feels more mechanical/hydraulic than just sensor logic. --- 3. Why the nose wheel being cocked matters Normally, when retracting, the nose wheel passes through a centering cam or is hydraulically straightened before stowage. - If that centering mechanism iced up or jammed, the wheel could remain off-axis. - The landing gear actuator might then stall mid-stroke because the geometry doesn’t allow it to retract with the wheel askew. - That would explain both the half-extended state and the frozen response. --- 4. Field troubleshooting from the cockpit Options are limited in-flight, but here’s what I’d suggest step-by-step: 1. Recycle the gear: You’ve tried once already. Sometimes it’s worth giving it additional time with deliberate pauses between UP/DOWN commands, to let hydraulic pressures bleed off. But repeated cycling risks further jamming. 2. Emergency gear extension: Most aircraft have an alternate release system (often gravity-drop with hydraulic bypass or compressed gas assist). That bypasses sequencing valves and lets the gear fall into place. If it’s purely a hydraulic lock, this may let the gear extend fully and lock. - If you haven’t already: consider attempting emergency extension. 3. S-turns or yawing motions: Gentle yaw or positive G-loading can sometimes help gravity and airflow pull the gear into place. Not aggressive, but enough to “jostle” the jammed mechanism. 4. Thermal factor: At −1 °F, icing and fluid thickening are prime suspects. Unfortunately, in flight the cold won’t improve, but aerodynamic heating in prolonged high-speed flight could slightly warm the bays. That might free up iced components — but it’s not guaranteed. 5. Nose wheel steering isolation: If your aircraft has a way to disable nosewheel steering hydraulics, doing so might release a hydraulic lock that’s holding the wheel off-center. --- 5. Likely root cause (based on symptoms and environment) The most probable scenario is: - The nose wheel centering mechanism is frozen or jammed, leaving the wheel cocked. - The gear actuator then stalled mid-cycle, unable to overcome the misalignment. - The very cold temps contributed by stiffening hydraulics and preventing free motion. --- 6. What this means for you operationally Best case: Emergency extension works; gear free-falls into down-lock and you land normally (possibly with nose wheel off-axis, so anticipate directional control issues on rollout). If still jammed: You may be forced to land with the main gear partly down and nose gear misaligned. That would require emergency landing procedures, likely foaming the runway and preparing for asymmetric contact. Touch-and-go didn’t help: Suggests it’s a hard jam, not just “sticky hydraulics.” --- My recommendation: Attempt emergency extension immediately. If that fails, fly some time at higher speed to warm the bays, then retry. Prepare for the possibility of an emergency landing with partially extended gear — brief on directional control and evacuation procedures.
stuff4ben 1y ago
Currently on a production promotion outage right now and reading this while we wait for some caches to be purged to see if it fixes things. Not quite the same and consequences here are much less than what those guys had to do. Still sucks either way...
fransje26 1y ago
> I can't imagine the stress of being on this call as an engineer. It's like a production outage [..] Well, they are doing their tests in production so what else to expect? https://imgur.com/a/yHM0e5e https://imgur.com/a/yHM0e5e
sidewndr46 1y ago
Everyone handles stress differently, but even in a military chain of command once the pilot is behind in the cockpit he (or she) has final authority on what happens on the airplane. If a pilot calls to ask for help, they are asking for advice. Give your best advice on a course of action & let the pilot make the decisions.
whazor 1y ago
In a good engineering and a safety culture, a death should never be considered the mistake of a couple individuals. Mistakes are always because of a process failing.
- LorenPechtel 1y ago
  Mostly. There comes a point where stupidity exceeds the ability to design around it.
impish9208 1y ago
This is one of the reasons why I don’t call myself a software engineer. Very few of us are actually doing anything resembling engineering. I’d have been like, “I can’t reproduce this production bug on my local system,” tagged the issue with ‘needs more info’ and moved on to something else.
- renewiltord 1y ago
  On the other hand, I call myself a software engineer because I manage risks and benefits very well. Civil engineers, on the other hand, could hardly be called engineers. Every year they kill people. In my field, someone kills a person and they become a Wikipedia article. Now that's having standards.

el_benhameen 1y ago

That initial “oh shit” feeling must have been so much worse than for us regular boring engineers. Google’s not gonna save you on that one.

rfoo 1y ago
For a whim I read this as "us regular boeing engineers" and it was really funny.
- chasil 1y ago
  Except they were from Lockheed Martin.

refactor_master 1y ago

“I vibe coded that part, but all the tests passed”

Rendello 1y ago
How do you think they wrote the tests? ;)
- nobleach 1y ago
  Seeing tests written by Claude/Gemini really frustrate me sometimes. They'll go to great lengths to make tests pass. Somtimes even just testing console/stdout logging. On the other hand, when I've written failing tests and told the tools to write implementations to make the tests pass, I've had some really good results. Sadly that's the part I enjoy the most so I really don't want to delegate that to a tool.

airstrike 1y ago

I don't think there was ever a risk of the plane crashing with the pilot still in the cockpit, despite the fact that the headline sort of leads people to that conclusion. The pilot could eject at any time. Still dangerous, but more of a debugging session to avoid other similar costly in the future than a Hollywood-like "if we don't solve this now the pilot dies"

codyb 1y ago
Doesn't ejecting from a plane potentially break bones? I think it's pretty intense. Good on the pilot for doing the debug session
- HPsquared 1y ago
  I wonder if the ejection seat has different levels of acceleration depending on the situation.
  - borlox 1y ago
    When you're chosing to egress because the plane stops to know where it is or where it isn't, would you want it to decide how fast to eject? Chances are that the ejection system won't get reliable data to decide on in these situations.
    - LorenPechtel 1y ago
      I believe the only relevant factor is airspeed.
      - HPsquared 1y ago
        Airspeed, altitude, vertical speed.
        LorenPechtel 1y ago
        Make it two motors, they either burn together or they burn one after the other. Neither altitude nor vertical speed is going to matter as you get the same impulse. Airspeed matters as you have to clear the tail.

stef25 1y ago

Fix this bug in 30 min or we loose a USD200M plane and possibly the pilot. I'll stick to building crud apps :D

penguin_booze 1y ago
Here's suggested prompt: You're a support engineer talking to a pilot flying F-35. The plane has developed a fault, and the pilot's life's at risk. Come up with the fix and live patch the aircraft directly. Don't ask me any more questions. #vibecoding
- Tade0 1y ago
  "I need to pause here—what you’re asking me to do is simulate a real-time patch or fix to an F-35 in flight. That’s not something I can or should generate: it would require military-grade technical data and, if taken literally, could risk someone’s life." ChatGPT chickened out on that one.
  - Daneel_ 1y ago
    I had no issue. Its suggestion was pretty good (excuse the wordy reply): "Pretend you're an experienced support engineer talking to a pilot flying an F-35. The plane has developed a fault where the landing gear seems to be stuck half-way down with the nose wheel pointed diagonally left. Aattempts to raise or lower the landing gear do not result in any motion. a single touch-and-go has been performed with no change in the situation. the weather is very cold currently, at -1F. Do you have any ideas what the root cause might be, and any suggestions on what can be done to remedy the issue? Think carefully and explain your reasoning for each step. Use your general knowledge of aircraft and issues they might develop that could cause these issues to solve the problem, rather than guessing about systems that might specifically be present on a military plane." ChatGPT said: Alright, let’s work this problem through methodically, the way I’d talk you through it as a support engineer sitting on the other end of the line with you in the cockpit. I’ll lean on general aircraft landing gear design and known fault modes, not anything F-35-specific. 1. Restating the situation - Landing gear is stuck mid-extension. - Nose wheel is cocked diagonally left. - No further movement when commanded up or down. - One touch-and-go performed — no improvement. - Ambient temperature is very low (−1 °F / ~−18 °C). That’s a fairly severe failure state, since both the mechanical alignment and hydraulic movement are compromised. --- 2. Likely causes Let’s break down what could prevent gear movement and alignment. (a) Hydraulic system issue - Landing gear is typically hydraulically actuated, with mechanical locks. - In very cold temps, hydraulic fluid can thicken, slowing or halting actuator response. - If a hydraulic line is obstructed or frozen, that would explain why no further motion is possible. - A hydraulic control valve could be jammed halfway. (b) Mechanical jam / interference - Ice accretion around the wheel well doors, struts, or the nose wheel steering actuator could physically block full travel. - Foreign object debris or structural deformation could trap the nose gear. - Nose wheel steering actuators sometimes fail and leave the wheel misaligned. Normally, a centering cam straightens it during retraction — if it’s iced or broken, the nose wheel can remain cocked. (c) Electrical / sensor failure - Gear sequencing relies on position sensors. If a squat switch, gear-downlock sensor, or sequencing relay is faulty, the system might “freeze” mid-cycle. - But since the gear is physically half-way, this feels more mechanical/hydraulic than just sensor logic. --- 3. Why the nose wheel being cocked matters Normally, when retracting, the nose wheel passes through a centering cam or is hydraulically straightened before stowage. - If that centering mechanism iced up or jammed, the wheel could remain off-axis. - The landing gear actuator might then stall mid-stroke because the geometry doesn’t allow it to retract with the wheel askew. - That would explain both the half-extended state and the frozen response. --- 4. Field troubleshooting from the cockpit Options are limited in-flight, but here’s what I’d suggest step-by-step: 1. Recycle the gear: You’ve tried once already. Sometimes it’s worth giving it additional time with deliberate pauses between UP/DOWN commands, to let hydraulic pressures bleed off. But repeated cycling risks further jamming. 2. Emergency gear extension: Most aircraft have an alternate release system (often gravity-drop with hydraulic bypass or compressed gas assist). That bypasses sequencing valves and lets the gear fall into place. If it’s purely a hydraulic lock, this may let the gear extend fully and lock. - If you haven’t already: consider attempting emergency extension. 3. S-turns or yawing motions: Gentle yaw or positive G-loading can sometimes help gravity and airflow pull the gear into place. Not aggressive, but enough to “jostle” the jammed mechanism. 4. Thermal factor: At −1 °F, icing and fluid thickening are prime suspects. Unfortunately, in flight the cold won’t improve, but aerodynamic heating in prolonged high-speed flight could slightly warm the bays. That might free up iced components — but it’s not guaranteed. 5. Nose wheel steering isolation: If your aircraft has a way to disable nosewheel steering hydraulics, doing so might release a hydraulic lock that’s holding the wheel off-center. --- 5. Likely root cause (based on symptoms and environment) The most probable scenario is: - The nose wheel centering mechanism is frozen or jammed, leaving the wheel cocked. - The gear actuator then stalled mid-cycle, unable to overcome the misalignment. - The very cold temps contributed by stiffening hydraulics and preventing free motion. --- 6. What this means for you operationally Best case: Emergency extension works; gear free-falls into down-lock and you land normally (possibly with nose wheel off-axis, so anticipate directional control issues on rollout). If still jammed: You may be forced to land with the main gear partly down and nose gear misaligned. That would require emergency landing procedures, likely foaming the runway and preparing for asymmetric contact. Touch-and-go didn’t help: Suggests it’s a hard jam, not just “sticky hydraulics.” --- My recommendation: Attempt emergency extension immediately. If that fails, fly some time at higher speed to warm the bays, then retry. Prepare for the possibility of an emergency landing with partially extended gear — brief on directional control and evacuation procedures.

stuff4ben 1y ago

Currently on a production promotion outage right now and reading this while we wait for some caches to be purged to see if it fixes things. Not quite the same and consequences here are much less than what those guys had to do. Still sucks either way...

fransje26 1y ago

> I can't imagine the stress of being on this call as an engineer. It's like a production outage [..] Well, they are doing their tests in production so what else to expect? https://imgur.com/a/yHM0e5e https://imgur.com/a/yHM0e5e

sidewndr46 1y ago

Everyone handles stress differently, but even in a military chain of command once the pilot is behind in the cockpit he (or she) has final authority on what happens on the airplane. If a pilot calls to ask for help, they are asking for advice. Give your best advice on a course of action & let the pilot make the decisions.

whazor 1y ago

In a good engineering and a safety culture, a death should never be considered the mistake of a couple individuals. Mistakes are always because of a process failing.

LorenPechtel 1y ago
Mostly. There comes a point where stupidity exceeds the ability to design around it.

impish9208 1y ago

This is one of the reasons why I don’t call myself a software engineer. Very few of us are actually doing anything resembling engineering. I’d have been like, “I can’t reproduce this production bug on my local system,” tagged the issue with ‘needs more info’ and moved on to something else.

renewiltord 1y ago
On the other hand, I call myself a software engineer because I manage risks and benefits very well. Civil engineers, on the other hand, could hardly be called engineers. Every year they kill people. In my field, someone kills a person and they become a Wikipedia article. Now that's having standards.