The real tragedy, from a scientific standpoint in the optical, is that we're entering a new world of wide-field and sensitive time-domain astronomy due to huge improvements in computing power and sensors at the same time that humanity is about to generate thousands of bright transients.

The only practical approach I can think of for combating the constellations is distributed arrays of wide-field transient-hunters. As long as the transient is common-mode across telescopes, then it is real. If there are three LSSTs, and not too many constellation-satellites, that might work.

I just noted, in passing, that the term for the satellite arrays appears to be "constellation"; that is a bummer in this context.