I'm pretty positive this drone wasn't using very many advanced algorithms. I mean, in the base case, you could easily spoof the unencrypted signal and try to force the drone to change directions. Unfortunately, this really only works when you're using Pseudorange measurements, as opposed to Carrier Phase. Moreover, if there was an INS on the drone as well, this interference problem would become rather trivial.
When using a GPS, if you notice large gaps where your data suddenly "jumps" from one location to another, it tends to be due to poor satellite geometry, i.e.: you have a cycle slip and you need to reconnect to the satellite and estimate the carrier ambiguity again before you can continue to use that satellite's measurements. Of course, this sort of thing is only detectable if you use Carrier Phase measurements, because Pseudorange can have an expected error of +/-10 metres. Now, suppose you're using Carrier Phase to detect cycle slips, and somebody manages to still spoof the signal. With an INS on board, you can still detect the direction of travel, as well as the acceleration at which you are traveling. If your INS does not agree with your GPS, then it would be assumed that the GPS is erroneous, and thus to ignore it until it starts agreeing again. There should be next to no excuse for somebody to take over a drone other than a mechanical malfunction or shoddy software.
Lastly, the idea that the military channel (aka P-code / encrypted channel) could easily be spoofed is ridiculous. GPS gives updates for satellite position approximately every 2 hours, if I remember correctly. That means you could effectively update the keys for the encryption every 2 hours, if you were so inclined. On top of that, the bandwidth of the P-code is much higher, and when interference/multipath/noise/spoofed signals try to mess with that, they end up having a far smaller effect on the signal than is given to consumer / civilian channels. Overall, the P-code is really hard to break, since encryption roll over can update very frequently, and noise and other interference can be mitigated quite well. It should also be noted that military based systems use very expensive receivers, often ones that don't track weak signals (that are more easily spoofed) and ones that can track multiple frequencies (L1, L2, L5, L1C, L2C, P-code) as well as multiple systems (GLONASS, GPS, Beidou/Compass, and more). Quite frankly, it's too expensive and too difficult to spoof that many signals and systems, since GLONASS runs on completely different frequencies, and doesn't even use the same type of signals (Amplitude Modulated GPS vs Frequency Modulated GLONASS). Overall, this article is FUD, and shows that the writer doesn't know jack lanterns about GPS.