Access.

I spent about 100 hours on a shitty 11" Celestron last year learning all sorts of things about space and optics and image processing. I'm hooked even though the results are terrible. I can rent time on some big scopes in high places, but the results are still subject to all of the artifacts and attenuation of the atmosphere.

Putting up a large number of reasonably capable space telescopes could be a tremendous boon to space education and even science. Look at atmospheric transmittance charts, there are entire areas of the spectrum that we can't even see from the ground. Having all of that research piled up behind billion dollar hardware pipelines doesn't foster a lot of attention or momentum.

Access for learning and education isn't going to ever pay for engineering and launch costs. Plus, if you're really interested in actual science, there's 30 years of Hubble images available to work through. Crowd-sourced science would be a lot more useful filtering through that back catalogue than coming up with anything new to look for that's visible by something with much less power than Hubble.

Stable Diffusion for amateur space-telescope astronomy.

Generally, not for cubesats though, but irrelevant to disposable mirrors

a bunch of them together might be "crowdsourced" weather satellites.