Shuttle flew on autopilot. The crew was largely there to supervise, at least in terms of the flying up and back down to landing. Automated systems actually flew the craft. The pilot was the emergency backup plan. So Buran wasn't all that much of a leap.

Shuttle's crew did the things the autopilot could not, all the docking, all the satellite grabbing, and certainly all the science. Buran's one flight was just an up-and-back proof of concept. So a crew wasn't needed.

Without a crew, the pilots' only job would have been to save themselves in various emergency scenarios should the systems fail. Take the pilots out and you don't have to train/study/prepare for those scenarios, at least not before the first flight. By removing the pilots, Russia reduced early program costs. It was a form of technical debt. They would eventually have to prep for humans, but at least they could get the craft flying first.

The point I was making was that you need lots of support equipment for humans, which must be hard to fit into a spacecraft and pay to launch even if you know you’re not going to need it.

It gives you more flexibility. You can fly with a crew when needed and without when not. You could also fly with a crew but no pilot.

It’s been speculated that if the damage to Columbia had been realized quickly, a rescue mission could have been launched and the crew returned safely on Atlantis. This would mean the destruction of Columbia even if it turned out the damage wasn’t that bad, because it couldn’t return without a crew. This is one random almost-real example of where this capability could be useful.

One obvious example is test flights. If a crewed craft can fly autonomously then it can be tested without risking those fragile, hard-to-copy meatbags. There will be an example of this next month when SpaceX launches the first test flight of the Dragon 2. Most if not all other manned spacecraft were tested this way too, aside from the Shuttle.

I understand that, but it seems to me that spacecraft are generally placed where you’re likely to not see flexibility because adding it is extremely expensive.

I don’t think adding autonomy is all that expensive. You need almost all of the stuff anyway, since no pilot can fly a spacecraft by the seat of their pants. As I mentioned in an edit (which you may not have seen when you wrote this), the Shuttle was unique or nearly so in not being crew-optional.

I was thinking of this from the other way around, of adding piloting capabilities to an autonomous mission; you now need to adding a bunch of specialized equipment to keep the meatbags inside the spacecraft happy. I’m not sure which way the Soviet shuttle was designed: primarily human-piloted, with optional autonomy added as an afterthought, or meant to be pilotless but with people were added later. It’s the latter case I was concerned about, but I realize now that you’re talking about the former.

I don’t think anyone has designed a pure cargo craft and then tacked on crew capability. The requirements for a crew are so big (you need a substantial pressurized volume, life support, the ability to renter and land, etc.) that I don’t think it would be practical.