Recovering old satellites for servicing isn't currently worthwhile, but it wasn't particularly relevant forty years ago either.
The downmass requirement was never about recovering old satellites. A new satellite is expensive, but after years of micrometeorites, irradiation, vacuum welding and a 500 degree temperature change every hour or so, it isn't worth much more than scrap value even if the design isn't thoroughly obsolete by that point.
There are two obvious exceptions to that - new satellites in the wrong orbit, and seventies era spy satellites (your own, not someone else's).
Westar 6 was a brand new satellite with a dud upper stage. It didn't really need to come back to earth - the most efficient fix would be sending a small tug to push it into the right orbit. A small tug was not immediately available but a shuttle was, so it got landed and relaunched.
The shuttle was designed a long time before digital cameras could match the quality of film. Spy satellites of that era had to drop small reentry capsules containing the exposed film. Once the film ran out within a few weeks of launch, the only option was letting all the expensive optics burn up in the atmosphere. Being able to cheaply recover the whole satellite would have been very useful.
If the shuttle had lived up to its promise as a cheap and reusable vehicle, it might have made sense to launch with one KH-9 in the payload bay and return with an empty one from the same orbit.
However, even that sort of satellite recovery wouldn't have been the best way to make use of downmass. Mount a large camera in the payload bay, possibly even permanently, and use the shuttle itself as the spy satellite. Note that satellite recovery only happened once, but Spacelab utilized the downmass capability 20-30 times depending on what you count.
As for the present day situation, retrieving old satellites remains largely pointless, and digital cameras have removed the spy satellite exception. Spacelab was replaced by delivering equipment to/from ISS. That does require some generic downmass capability, but not a vehicle as big as the shuttle.
Recovering Hubble would be bringing back a museum piece, just because we can - I certainly wouldn't expect it to be worth relaunching thirty year old hardware.