JWST is for researching high-redshift object - the most distant things we can see in the universe. It's not a planet-hunter. And it will be much further away than other satellites, so they would be more easily serviced. (Not that it's easy, I just think you picked a bad example.)
> JWST is for researching high-redshift object - the most distant things we can see in the universe. It's not a planet-hunter.
JWST is a general purpose observatory. It will make significant advances for high-redshift objects. It will not survey large numbers of star for new planets, but it will be able to observe and characterize planets that have been identified by other missions/observatories.
SpaceX is working on Starship, which should be capable of Shuttle-like service missions. (JWST is much more fragile than Hubble, however, so you'd need to build a custom servicing solution... but JWST is not representative of space telescopes generally.) In fact, SpaceX is counting on Starship to deploy their full constellation. No Starship, no 12,000 satellites.
I think JWST is going to look a lot like the future of space telescopes. There's hard requirements that make larger satellites better for observation, principally because larger aperture == better resolution and also because the larger mirror gathers more light so we can look at dimmer objects.
It's not the size of JWST, it's the fact that it's deeply cryogenic and uses enormous, fragile multilayer insulation to remain so.
Indeed, larger apertures will be used increasingly. And this megaconstellation actually enables that indirectly. SpaceX is counting on their Starship launch vehicle to deploy the vast majority of their 12,000 satellite constellation, and indeed large launch vehicles (particularly reusable ones) require a lot of payloads to justify their development cost. But large launch vehicles also enable ever-larger telescopes, whose cost is only affordable if those launch vehicles are reusable.
NASA is already buying into it. Here's NASA Goddard tweeting out the larger-than-JWST concept called LUVOIR inside SpaceX's Starship (which again relies on the megaconstellation for demand): https://twitter.com/nasagoddard/status/1116310431969239040?l...
...and yet larger apertures will require in-space assembly over several launches, again only really affordable if there's enough commercial launch demand (i.e. from megaconstellations in particular) to justify these large reusable launch vehicles.
My point of view is that if we're going to be a really spacefaring civilization, it's reasonable that you'll be able to see human artifacts in the sky. Starlink is actually fairly easily stealthed, but other things (space based solar power stations, large space stations, in-space manufacturing depots, etc) may not be so easily hidden. Is the fact that human artifacts may be detectable in the night sky reason enough to disallow spacefaring?
EDIT: Note that the IAU letter is way more reasonable than almost all the media takes on this issue.
Your comment assumes we still need fleshy meatbags wearing clunky suits in orbit to fix things. This was true 15 years ago, but I'm not convinced it's still true today. A couple of repair drones housed at the ISS combined with repair parts brought up along with normal resupply missions could probably be made to work if it was deemed cost effective.
The ISS isn't in a particularly good orbit because it's highly inclined (so Russia could easily launch to it) which is an expensive issue to fix dV wise unless all the target satellites are in a similarly inclined orbit. In terms of making them easy to swap out that's certainly possible but it will add weight compared to a monolithic satellite and so far we have very very little experience building and repairing things in space without using people.