In defense of Elon, the Starlink satellites are equipped to de-orbit themselves at their end of life. They carry fuel for just that purpose.
Still, it remains a contentious issue.
Sadly many geostationary satellites that had such provision failed their deorbiting and are now cluttering that coveted high altitude ring, mostly forever since there is no abrasion due to the atmosphere up there. The saving grace of Starlink is that they operate at LEO (low Earth orbit) of 500 km and would eventually deorbit (aka burn in the atmosphere) by themselves in a few years.
Let's do the math: 12'000 satellites with a lifespan of 5 years, assuming a conservative deorbiting failure of 2 percent, would lead to 240 new dead bricks at LEO every 5 years. Assuming a reentry time of 10 years, there would never be more than about 500 dead bricks in the sky.
But if Starlink is successful, there will surely be competitors. China will want that, too, and what about Europe? Still not fed up with the Galileo sats clock failures? So let's multiply this by 4, 2'000 dead bricks at any given time.
Shall we stop here, before this gets boring?
As for photographers and scientists, a clever person will soon invent a pair of glasses with
augmented diminished reality that will filter the sat streaks. Same for photos and video clips: Topaz Labs will release a new AI tool to (mostly) remove the ugly artefacts. Sure, it might make existing workflows more complicated, but at least everyone on Earth will be able to download cat videos in 8K.
At least the few percents that can afford it. Because that is the absurd thing in the end: the whole of humanity will have to deal with Starlink pollution even if only a minority will use the service.
Cheers,
Fabien