> Moreover, in refueling ISS you can use something like a flexible bladder and pressure differential to simplify the job of moving liquid from container A into container B. But in the rocket-to-rocket case, you might be moving propellant from an almost-empty Starship into an almost-full depot rocket. In that case, you're trying to hunt around for liquid in an almost empty fuel tank, and push it into an almost full one.
> You can't use a bladder because Starship is too big, and it's hard to maintain a big pressure difference (unless you're willing to vent a lot of propellant in the process).
You just put a baffle in the tank so the volume with remaining propellant is small and close to full. Also eliminates sloshing issues that you'd need to deal with anyways.
In orbit fluid transfer is a reasonably solved problem, and there are many ways to do it. In addition to using bladders, you can also use ullage motors, centrifugal propellant settling, capillary tubes, etc. Cryogenics are harder because cryogenic pumps are just generally more challenging than standard pumps, but luckily the pumping requirements for propellant transfer are much less demanding than for engine restart - rocket engines need high flow rates and can't tolerate entrained gas, propellant transfer can use slower but more robust pumps.
I don't mean this to diminish the accomplishment of the engineers who spent quite a bit of time solving this problem, but rather to point out that people have been working on this for decades and have had significant success. Sure we haven't done a perfect 1:1 dress rehearsal in orbit where large quantities of cryogenic propellants are transferred between docked spacecraft, but we've done everything shy of that.