...distort? You mean the frequency curve inaudibly dips, not actual distortion.
This is really academic. No one is going to hear the tiny dip in the O2's driving of a low-impedance balanced armature. The design principle was to aim for audible transparency (which pretty much defines true hifi) not technical perfection, which is in anycase impossible without a true Zero Z (output impedance) which is impossible if I remember correctly.
Technically that's distortion, but almost everybody means "non-linear distortion" when they talk about distortion. Mild FR changes are not non-linear distortion, as you say. I wouldn't be confident enough to say "inaudibly dips", but I think it's at the level where it's not worth losing sleep over. That said, it's a bigger deal than some other things that audiophiles worry about.
Many devices have lower output impedance. 0 output impedance is like 0 distortion or 0 noise—impossible, as you say... unless you find me a wire with 0 resistance and much more. It's meaningless in the grand scheme of things. Most of it on the O2 comes from the output resistors, which are there mainly for safety / matching of the cheap output op amps. A small tradeoff, like many things.
If you wanted, you could adjust the gain on lots of other DIY amps, some commercial ones too. Though beware of instability.
