The probability of something big hitting us is really low, coupled with the fact that we can see it coming since all objects in space are visible. Miss-calculating a trajectory could indeed happen and we may not notice it if it's coming from the direction of our sun, but the probability of that is even lower.
So why would you spend tons of natural resources, time and money for a machinery of dealing with a problem that we may not have in the next 100,000 years? And what should that machinery do in the meantime?
Personally, I think our world has to solve bigger and more imminent problems that can lead to extinction too ... hunger, disease, overpopulation, economic recessions, natural resources running out, global warming, the destruction of the only habitat we know that can sustain life, nuclear wars and the list of things that can happen in the next 400 years can probably go on. Heck, we came closer to extinction during the Cold War than we ever ever did from a comet passing by and the only thing saving us was the few reasonable men in charge, but we may not get so lucky next time.
The majority of objects in space are not visible to us. They only become visible when near and bright enough. Even known comets such as Halley are not visible to us most of the time.
I don't know much on this topic, but if we are able to spot planets that are light years away, then what's the problem with detecting objects that are gonna crash into us in the next 1000 years?
The only problem I can understand is that the sky is a pretty big place, but surely we also have a lot of time to detect such objects.
Never mind the fact that even if we have a telescope that big, the issue is that these telescopes show a microscopically small field of view. There's a whole sphere that would need to be observed, and repeatedly (since the bodies move).
I'm no expert, but as I understand it, the kind of objects we would be worried about are extremely hard to detect as they are dark, cold and relatively tiny.
At the same time, our "eyes" can only look at tiny patches of sky, generally through the distortion of our atmosphere for only a portion of each day with parts of the sky only visible part of the year. Furthermore, we generally have those eyes trained to other, known targets - stars, galaxies, nebulae and all that.
Consider that we didn't spot Apophis [1], which is set to come extremely close, until 2004 even though it's on our lawn [2] in terms of distance.
There is an effort [3] or two [4] afoot related to tracking asteroids and a new DARPA [5] program for tracking our own local junk.
1: http://en.wikipedia.org/wiki/99942_Apophis
2: http://www.howcloseisapophis.com/
3: http://www.skyandtelescope.com/news/160839035.html
I am not sure if this is in line with Agile principles like YAGNI.
It's hard to estimate how long it takes for another humanity to develop if we screw it(so we don't know iteration length), and there is the problem of telling them what we did wrong after we were destroyed.
There needs to be time for technology development, deployment, and testing. Tackling an extinction-level event is not something we want to scramble for at the last possible minute, using untried technology.