„In weniger als einem […] Jahr gingen die Teams von der Erstinstallation der Live-KI an […] in den X-62A[…] um die erste KI im Vergleich zum Menschen zu demonstrieren [in] ein Luftkampf.“ – USAF Test Pilot School und DARPA geben Durchbruch beim maschinellen Lernen in der Luft- und Raumfahrt bekannt (Chase Kohler, 2024)
https://www.afrl.af.mil/News/Article-Display/Article/3745032/usaf-test-pilot-school-and-darpa-announce-breakthrough-in-aerospace-machine-lea/
1 Comment
__Submission statement__: This was to be expected, given the existence of [DARPA AlphaDogfight](https://en.wikipedia.org/wiki/DARPA_AlphaDogfight), though in this case the system hasn’t been trained via self-play, but on human dogfights. It’s relevant because ditching human pilots probably can enable a much wider range of maneuvers—fleshbags have a problem with high [__g__](https://en.wikipedia.org/wiki/Gravitational_acceleration). Curious whether there’s similar efforts in the PRC (or other countries), and how they will play out in relation to drone warfare.