I more-or-less agree with this much, but the most suitable place to start learning to colonize space rocks is probably Phobos.
Most worlds in our system don't have a significant atmosphere, most of those that do have too much of it. Like gravity, atmosphere also poses a problem for getting back to space.
> no man-made spaceship can match that for billions of years.
Not unmaintained, no, but if it was unmaintained that probably means everyone who lives there is dead anyway.
I always believed that if humanity is to colonize something, it would be through robots that do all the exploring and mining for us.
Yes, robots are a cheaper and more reliable way to explore and extract resources but I doubt mankind will be happy with just that. We like to explore, and expand, and face challenges so I'd bet that no matter the setbacks or price tag, we will eventually prefer to do this things in person.
As an alternative: we could even redesign ourselves for space exploration if needed be.
As to your wider point, the ability to breath oxygen and have long term fat stores makes an Atlantic crossing very easy. We moved from one palace with humans to another place with humans, hardly a massive feat of engendering. Meanwhile people in far more primitive craft ended up living in Hawaii (2,200+ miles from the nearest land mass) of all palaces.
Which is exactly why low G-tolerance doesn't matter. You have a limited amount of delta-V, whether you apply it all in one go or spread out over a year makes little difference given the timescales that are already involved in interstellar travel.
If you give a mars rover a command, it takes an average of 12 minutes for that command to reach mars, and another 12 minutes for the confirmation to come back. That makes remote control very hard, and we aren't all that good with autonomous robots.
So until we figure that whole artificial intelligence thing out humans are a much better bet for getting significant amounts of science and mining done than robots.