It's a bit hard to visualize because we're used to things like heat shimmer, fog, the curvature of the earth and the very air itself blocking our view. Actually that's what those observatories have to keep in mind as well, but only about 100km of it instead of the distance as mentioned in the article.

It's not the size of something that makes it visible, but its brightness. Without actually doing the calculations, I'd bet there are stars you can see with your naked eye that are similar in apparent size or even smaller. In order to see detail, you'd need a lot of magnification. But your eye or a camera are able to detect that "something" is there due to the fact that there a lot of photons coming from that direction. The article talks about them observing dips in brightness, so they were probably just looking at a single pixel, or a handful of pixels at best.

They described seeing a change in position at that scale. Brightness is one thing, but detecting an angular change like that is mind boggling.

It feels like you'd have to start compensating for things like plate tectonics and daily variation in the earths rotation due to melting glaciers to get that kind of accuracy.

The question in my mind is how can you point a telescope when the accuracy is that high. A heavy truck parked 100 meters away probably deforms the ground locally enough to change things.

Steel probably wobbles like jello when motion at those scales matters.