"model failures become increasingly dominated by incoherence rather than systematic misalignment."

This should not be surprising.

Systematic misalignment, i.e., bias, is still coherent and rational, if it is to be systematic. This would require that AI reason, but AI does not reason (let alone think), it does not do inference.

A thread on beating EW against Starlink. https://x.com/TrentTelenko/status/2011251398055489972

Couple of tips: burying Starlink in a hole a meter below ground level blocks ground-based jammers.

Mesh nets around the antenna to block signals, effectively a high-pass filter.