Hacker News new | ask | show | jobs
by whartung 5 days ago
As I understand it, they done this in simulation, and the human pilots did not fare well.

I don't know of the computer aircraft were able to fly beyond human limits (i.e. they don't black out) or not. If the computers can actually fly to what the airframe is capable of, they would have a distinct advantage over the human pilots.

2 comments

They did a small set of tests with one Pilot IIRC?
I think they did it in simulation with several.

In any case, dogfighting is commonly done because it's great training but in most cases, you expect most Air to Air combat being flinging missiles at each other.

s/most/all/g
Exactly.
The simulator is not the aircraft.
The first time an airline pilot flies a new aircraft type in the real world they fly it with a plane full of passengers. All the training up to that point is done in a simulator.

Now that first flight is with an experienced training captain but the point is the simulator is good enough that the FAA has signed off on using that as the primary training device.

It's been this way for at least 25 years.

Many of the crash reports from the likes of Mentor Pilot on YouTube call out insufficient type-experience as a cause of accidents. Sometimes it's "the same" type, but the manufacturers manipulated the situation, the changes mean unfamiliarity and that leads to deaths.

Not giving pilots live training on new types is for financial reasons. It doesn't cause more deaths than the cost differential between training and compensation allow for.

no airline will take the risk. they will have a check ride, with an instructor.
Simulator tech has come a long way. I wouldn't say a simulator invalidates the result.
with respect to assessing tactics, it is very close.
With respect to assessing whether an AI-controlled aircraft using only available onboard sensors and datalinks can defeat a manned aircraft in BFM, it's not even in the same ZIP code. You've never flown BFM, have you? I have.
I yield to your experience.