I would also argue that life itself has proven remarkably long-term stable on our planet. Granted, it has not been stable enough to produce a type-III society, but if we remove homo sapiens from the planet for a thought experiment, there exist other species that are developing communication or tool-use skills and might succeed us even on the technological path.
So my view is that, unless we continue to fail spectacularly, life will still continue without us. Even if we do not reach type-III, that does not necessarily preclude Earth from producing a type-III society in the future.
Sounds like you are the one who has a narrow view of "society," rather. In the abstract, it only need be something like "a multitude of cooperating but competing, independent self-preserving, self-advocating sentient interests." That's fairly abstract while still being a sure recipe for highly non-linear and chaotic effects.
Even if we do not reach type-III, that does not necessarily preclude Earth from producing a type-III society in the future.
But then you're setting up a situation where a series of chaotic and unstable arrangements snuff themselves out or flare-out in a puff of singularity-colored smoke, ending with the appearance of the first stable meta-societal form.
"all" is usually the key issue with solutions to the Fermi paradox. We can certainly find reasons for any given civilization to go the other way, but it's more difficult to argue that all of them do.