Well... let's not just brush it off as one single management failure with no precedence.
Note that even the HST was hugely over budget, over time, and was mismanaged as well... so it's not like it's a one-time thing. Instead, it's more common than not.
From its original total cost estimate of about US$400 million, the hubble space telescope cost about US$4.7 billion by the time of its launch. Hubble's cumulative costs were estimated to be about US$10 billion in 2010. It was launched years behind shedule. It had a flawed main mirror. Etc.
So...sure, it provided nice pictures, and was/is a great asset, but let's not sugarcoat the truth: it was over-budget, over-time, and mismanaged as well. You could have built about 3 OWL's with that. And those would have provided better pictures, certainly in the visible spectrum.
That's why the "Just because in e.g. 20 years we can do something much cheaper we should not stop building scientific instruments today." argument doesn't hold up, imho, because it will ALWAYS be cheaper to build them on Earth. It's a given. One knows this in front. It's not about 'not' building scientific instruments; one can build new ones for Earth-based telescopes as well. It's balancing the advantages and deciding where the most bang for the buck is.
And as said, I think in many instances, unless for specific reasons or goals, space-based telescopes will always come out lower in that regard.
Now, I do understand what you're trying to say, and it's one of the reasons why - even though I personally would have liked to see 'normal' pictures - I do think it made the most scientific sense to use JWST for infra-red observations. At least, that's something that can not, or only with great difficulty, be done on Earth. But still, it's hard to argue that it was worth more than 10 billion dollars. And it's also hard to argue that's an exception, in large space projects.
with the money of the HST, one could have build a couple of EEVLT instead. With the money of the JWST, one could have build a couple of OWL's. And if one ever would send up a new teslecope, double as grand as JWST and double as costly, it's a given for the same money one would, once again, be able to build a far more powerful telescope on Earth. Yes, maybe not in the infra-red... but is getting infra-red really worth 20 billion dollars?
As for exposure-times... you could build an OWL on each side of the planet. Also: while exposure time is important for deep field viewing, *aperture* (and thus lightgathering power) is even more important. Meaning: with a much bigger mirror, you can see much more much faster. So one day with a two meter diameter telescope would demand LESS than half a day with a 8 meter telescope, if we're only talking about exposure time.
And sure, space has its advantages, but the point I'm trying to make, they're not all THAT overwhelmingly large anymore, on a lot of fronts. Not to warrant a cost/time loss of billions/years - and certainly not if it's not strictly necessary to do it that way.
Now... true, costs may come down - hopefully that will pan out for SpaceX - but if we're speculating on future technology, one may do it both ways, and what if, in the future, they find a way to capture infra-red waves in sufficient amount on Earth, for instance? Future technology works both ways, after all - adaptive optics, unimaginable only 40 years ago, is the proof of that.
So, while potentially complementary, I think a very stringent look is necessary as for what projects, exactly, a space-telesope is worth the extra effort and money, and to what degree.
For some tings, it's pretty obvious; for instance, if one wants to test out interferometry on a scale larger than the diameter of Earth. For others, it's less obvious or necessary.