I’m inclined to side with the “we know almost nothing” side of the argument. The problem is that our sample size from which to draw conclusions is hopelessly small, so we have to make a lot of fanciful assumptions, which may well be completely wrong.
Just about the only assumption in which we can have some level of confidence – though that might be wrong, too – is the assumption that complex life, or indeed probably any life at all – must necessarily be based on carbon. That puts constraints on the conditions under which life could develop and flourish. But there’s very little else that we can assume with any confidence.
Our search for life thus far has been very local and based on incredibly limited evidence. Within the solar system, only Mars has been explored to any meaningful extent, and other candidates for life like the oceans of Europa and Enceladus are a complete wildcard. On a larger scale, our efforts to detect signs of distant intelligent civilizations elsewhere in the galaxy are based on highly conjectural approaches and, moreover, are also intrinsically local. This is because our galaxy has a radius of nearly 53,000 light years, and we’re located in quite a remote part of it. The “downtown” part where the action is is clustered around the center, and that’s around 26,000 light years away.
These vast distances not only make any signs of other civilizations almost impossibly hard to detect even if we knew what to look for, but it also introduces temporal factors. What was the state of our own technological civilization a mere thousand years ago? When did we first start broadcasting potentially detectable radio waves, and how much longer will we continue to do so? And, indeed, the development of science and technology was arguably a total fluke; we lived for tens of thousands of years in more or less our present form and developed virtually nothing at all, and may well have continued that way for thousands more, but for a quirk of fate that started the ball rolling on the exponential growth of scientific knowledge and its applications.
All of which means that if we expect to detect signs of distant civilizations, we not only have to be looking at exactly the right place in the unbelievable vastness of space, but we have to be looking at exactly the right time, because the signals may last much less than the thousands of years they took to get to us before the civilization either died out or moved on to different endeavors, and that’s assuming such signals are even theoretically detectable against the background noise of the universe.
As for “Dyson spheres” and the like, any notions about what sorts of technologies super-advanced civilizations might create is so speculative that it doesn’t even rise to the level of scientific hypotheses, even though it’s often dressed up that way. It’s pretty much science fiction.