This is not a device that will be produced in an assembly line. Efforts to improve its production would be nullified by the fact that every unit is itself a prototype.
I wonder about that. Hubble, after all, could have been looked at the same way from the outside, but it was essentially a proof of concept for a constellation of top secret satellites still in use today.
The opposite is true. Hubble is a stripped down version of a Keyhole spy satellite that was already in production. One of the major repairs (replace solar panels) was required because of long duration exposures used in astronomy that are not used in spying.
Remember that Hubble was essentially useless until heroically repaired the first time by Story Musgrave. Many subsequent repairs were required to get the best pictures from it - like the deep field, long duration picture.
There's another way to think of this. Let's use Starlink as an example, which is ~1300 today growing to 40k satellites in LEO.
If each of these had a high-grade commercial camera lens pointed towards earth, you could have real-time viewing of the entire planet with just a little bit of math. With a bit more fun (and for each click of optics improvements), you could "Bullet Time" view a great number of things...
If each of these had a high-grade commercial lens pointing outwards, what could that be used to capture? At scale, it seems like doing diff's against all the images would reveal a huge number of previously unidentified asteroids and other near-earth objects.
What if you put "very good" optics on them? Or a radio antenna pointed out? Could you make earth sized optical or radio telescopes? Seems like you could.
It's increasingly looking like the assumption each one needs to drive the state of the art, and each one needs to be "space grade at 100x the cost" is a fallacy simply because launching was historically so expensive.
Very frequently in this type of conversations a layman suggests "just make a constellation of low/average quality telescopes that acts like one big telescope". Presumably because they have heard of "long baseline telescopes"[1]. No, that does not work in the optical and infrared because you can not keep the signals from different telescopes coherent with one another (the technology to do that is science fiction today and it would be crazy expensive the first time realized). This is not to say there is no merit in the idea that many cheap telescopes can be useful in some other (non-visible-or-IR-spectrum) way, but the typical bombastic suggestions are simply wrong and are a disservice to an otherwise interesting idea. We will continue needing big-mirror optical and IR telescopes for quite a while if we want to continue studying the universe.
For the cameras facing earth, you wouldn't need fantastic resolution for it to be useful. With a resolution of 50 meters, you could track suezmax ships. With a resolution of 5 meters, you could track commercial aircraft. Not great, but it would help with a Malaysia Airlines Flight 370 like incident.
Biggest issue is that the efficiency of telescopes at detecting faint point sources scales as (mirror diameter)^4th power -- strongly favors large telescopes over arrays of small ones.
With asteroids, you don't need to diff telescopes at different locations; you can just diff the image taken at one location, 15 minutes apart.
