I can't find the source now, but some estimate that removing only a few of the largest boosters on orbit each year would decrease the cascade risk significantly, since those boosters represent a large fraction of the mass. This is a much smaller move than many other avoid-the-end-of-the-world suggestions given throughout the years.

Now: better modeling would probably help us determine whether Kessler syndrome will actually happen.