It may not be exactly either/or, but budget constraints coupled with the astronomical cost of using humans in space makes it nearly so.

I think the example of Hubble actually works in my favor on this point. Hubble is a relic of an old technological age. Earth-based telescopes surpassed Hubble's capabilities not too long after it was launched, and have now pretty much left it in the dust. At a fraction of the cost.

Hubble's replacement (James Webb) is quite a leap ahead, and as an IR scope it really does need to be in space, unlike Hubble. But you'll note that its location will be off-limits to humans, and it is accordingly designed to be remotely maintained.

In short, the age of low-earth-orbit, high-maintenance space telescopes is, and should be, over. I argue that other space-based platforms that require human intervention should likewise be curtailed for similar reasons.

>"Automation isn't cheap, missions that have to be redone because the automation fails are costly."

It is still orders of magnitude cheaper than sending people, and the cost of a "failed" space endeavour with people on board is infinitely greater. In cases where it's not currently feasible to use robots, it would be worth our while to wait until it is, instead of trying to use humans now.

I feel that Mars is another example in my favor. Cost estimates for sending humans to Mars are so high and so speculative that they're essentially useless outside of the insight that we cannot afford it. Compare that with the cost of robotic exploration, which is both quite affordable and spectacularly successful so far.