I've always wondered whether payloads weren't super-expensive precisely because the launches were, too. I mean, if you are spending hundreds of millions to lau

I've always wondered whether payloads weren't super-expensive precisely because the launches were, too. I mean, if you are spending hundreds of millions to launch something, that something better work once up there, right? So you spend 20 years engineering a "perfect" payload, the costs do what they always do, and now you have a billion-dollar payload. But if a launch was $2M and bookable with a month's lead time, you'd just de-risk the whole thing by having a simple-to-build, likely-to-work payload. Or ten. (This might not apply to something like the JWST specifically, since even a terrestial mirror of that size is a huge challenge. But I bet a lot of other kinds of science payloads will be built very differently a decade from now.)