Automation is especially challenged by richly varying or adversarial conditions. The moon has much less of both than the Earth. I already agreed that the particular conditions include new problems; in fact I already listed your first two.

BTW spacesuits could probably be much better for repair work; they seem like another area where NASA has stagnated.

On Earth you need to compete against other people doing the same. So you design on the edge of performance to extract the last few percents of efficiency to compete on price against all the other people doing the same thing. Which means the machines are complicated, use rare materials and require a lot of maintenance.

On the Moon you can do the simplest thing that works and if it works at 10% efficiency and breaks after 1 year - so be it, if it's enough time to get resources to make a new one.

Basically space exploration will have a lot more in common with industrial revolution than with overengineered spacematerial NASA stuff.

If we have to make the tractors 10x bigger to have the same power and output, and to use disposable steel cables instead of hydraulics, and to make them disposable after 2 years instead of lubricating them to last 20 years - that's all fine if it means it can work with lunar materials only.

the size. An optical telescope is limited by the size of the mirror which needs to be one continuous surface (though not necessarily smooth for instance the JWT but I digress), but a radio telescope doesnt, it can be many individual collectors that can be joined together. This enables it to be much larger so it can collect more signal, and the radio waves are much longer so it needs to be much larger. In an extreme example we have used many different radio telescopes together with very precise timing to produce the images of the black holes at the center of M87 and the milky way.

But it requires those different clusters of collectors to be stationary - so while you could probably build a swarm of satellites, they would have to stay in very precise distances from each other over time which would be considerably more difficult than planting them on a surface.

Also, a big shield like the moon blocking out radio interference coming from the earth is desirable.

IANAA, corrections to my understandings welcome

It’s difficult to keep things spaced out properly in microgravity without scaffolding. See also https://en.m.wikipedia.org/wiki/Project_West_Ford

I guess local resources are easier to use locally, rather than launching to orbit first? You also have a firm base to mount the antenna, and the process of mounting could be arguably easier.

They're planning to build this telescope on the Moon, by mining Moon rocks.

In space, there are no rocks to mine, so you're have to launch all the material to space, which is wildly impractical/expensive.

Easier to build it where the materials are than try to launch a city-sized radio telescope.

I think humans currently have a hard time deploying a 20km by 20km object in space.