Correct, article says they landed with 220kg which is around 6 minutes of average fuel burn over an entire flight - bit less at cruise, a hell of a lot more at takeoff/climb.

So I don't think 220kg is enough to do a go-around in a 737 (well, a go-around would've been initiated with a bit more than 220kg in the tank - they burned some taxing to the gate - but you get my point.) I've read around 2,300kg for takeoff and climb on a normal flight in a 737-8. A go-around is going to use close to that, it's a full power takeoff but a much shorter climb phase up to whatever procedure is set for the airport and then what ATC tells you.

I just flew 172s but even with those little things we were told, your reserve is never to be used.

These people came very, very close to a disaster. Fortunately they had as much luck left as they did fuel.

> I've read around 2,300kg for takeoff and climb on a normal flight in a 737-8.

ChatGPT suggests way lower numbers for take off / climb in a 737-8.

600-1000kg from brake to reaching cruise.

That’s about as useful as opening a fortune cookie and reading it off as an answer.

Straight from the horse’s mouth: https://web.archive.org/web/20230630013840/http://www.boeing...

In the first table they list 2307-2374 kg of fuel for takeoff and climb.

You’re talking to the wrong horse though.

Isn’t a 737-8 the max 8 variant? It uses newer dual CFM LEAP-1B engines. How does it compare? I can’t really find the data. The spec you’re referring to is for the older 737-800.

Another fortune cookie:

https://www.aircraft-commerce.com/wp-content/uploads/aircraf...

It suggests an overall savings of ~14% over the 737-800 but doesn’t look at specific takeoff/climb comparison.

I wasn’t posting the LLM output as a source of truth. I was just using it to question the uncited value. And I still really don’t know the answer. If you’ve got another data source I’d love to get it.

ok, how do we verify that?

Maybe he should ask Claude next.

That’s the point? I wasn’t suggesting it was correct. Just that the value is wildly different from their own non-cited number. The next stage was to get a citation from an actual datasheet. Their reasoning was nothing beyond “I’ve read”

Why do people keep insisting on pasting LLM output to HN when every time it happens, it gets downvoted to oblivion? The community clearly doesn't want it. If we wanted to know a computer program's opinion about something, we could ask it ourselves.

I was using it to question that exact stated fuel consumption number without a citation. For hard data (like fuel consumption) getting a value from an LLM isn’t absurd.

If not absurd, it's very poor form. You should never use LLM as input for a discussion, nobody wants to hear that. Use it to search for authoritative sources.

It’s fine if you post an actual citation that you might have found through the LLM. Just posting AI slop is worse than useless, though, and also unpleasantly dystopian.

I agree, well out of comfort zones. However to my reading multiple different things went wrong to get to this point.

That could be. We just don't know right now, but your intuition may well be correct, even if there is a single root cause there could very well be multiple contributory causes.

They failed to land at two airports before the third. I can't say if they made the right decisions but that already is two failures.

Go arounds are not failures.

They are expected situations, but still a failure of the original plan.

They are not a failure of the original plan, they are a mandatory component of the original plan that if everything is nominal never gets executed. Every pilot on approach is ready for one or even more go-arounds and they happen quite frequently for a variety of reasons.

They happen a few hundred times per day at ~100 k flights.