So my proposal is this: Spend a lot less money building a much larger, much more durable self-sufficient colony of robots. Once that's going good and strong, maybe we program the robots to produce the oxygen and water and complex organic molecules and radioactive shielding we need in a reliable enough way that it's worth sending people. If that all doesn't quite work out, well, we still got a self-sufficient colony of robots.

We can't even put a self-sufficient colony of people in Antarctica, where we don't have to worry about the lack of atmosphere or radiation shielding.

No, we haven't attempted to put a self-sufficient colony of people in Antarctica. Which is not at all the same as "we can't".

Right. We attempted to put a self-sufficient colony in the American Southwest and still failed.

Surely it's fair to extrapolate our inability to succeed there into an inability to succeed in a far more harsh and remote locale.

Keep in mind that this isn't to say we could never succeed. Just that this is a very, very difficult problem that we don't even talk about seriously today, let alone study and pursue. Yet everyone just assumes it's somehow trivial and automatic once we decide to try.

Assuming you're talking about Biosphere 2, do you know why it didn't succeed?

What problem did it have that is unsolvable?

Pauly Shore.

Antarctic colonies haven't needed to be self-sufficient, given the relative ease of resupply, so they haven't been.