So, it seems a bit disingenuous for the article to make a big deal of LEO satellites.
The assumption I think you're making is all debris remains in a fairly localized decaying orbit, which is not the case when a high energy collision occurs. The debris will fan out in two radial smearing patterns that loop back around at the collision point. Some pieces will hit more atmo for longer periods while others will achieve higher altitudes for a time and those can debris fields last longer. Should a LEO cascading collision event happen we would see lots of debris fields reaching higher orbits and a non-zero risk of those satellites being hit too.
* we do not have any kind of technology to get stuff back from orbit. Anything that is there and anything that we put into orbit will stay there, passively, until the laws of physics take care, the only exception being satellites that are made to plunge back into the atmosphere with their left over propellant. We cannot even hope to get meaningful amounts of plastics back from the surface of the oceans which is a similar problem only many, many orders of magnitude more feasible.
* When you put satellites into lower orbits you'll need more of them to obtain the same coverage of ground. At the same time, the volume of space is smaller in lower than in higher orbits, so more crowding. You can use smaller, less powerful satellites, but because of atmospheric drag, their lifetimes will presumably also more limited in low orbits, meaning considerably more additional traffic due to the increased need of de-orbiting and replenishing activities.
* Should anything go wrong (and something will go wrong sooner or later, as indicated by the fact that we already have to close keep tabs on the course of the ISS), a collision will spread parts into all directions. One would imagine that in LEO roughly half of the sky is occupied by nearby Earth, but the other half is pointing sideways or up from the Earth, so if two 100kg satellites smash into each other you still have like ~100kg of new debris that will fly into directions not prone to natural fast decay. Of course this is just a rough estimate on which I'd like to be proven wrong.
* Low orbit and increased number of objects also mean more interference with astronomy. I cannot believe when folks claim that now that we 'have AI' we can just 'use AI to clean up astronomical imagery'. Our entire ability to put stuff (and people) into space at all is predicated on many, many millennia of gazing up to the skies and trying to figure things out. No astronomy, no going to space. When you take an image and there's an object blocking the view, or worse adding a very considerable amount of reflective light that saturates your sensors, no amount of 'AI' will get back that info to you because the data is, essentially, just not there.