Evolution is a poorly teached subject
I saw an op-ed piece about the pseudoscience status of intelligent design lately on kuro5hin.org, and I must say that it immediately got me down mostly because it reminded me how poorly evolution is taught on schools, and how superficial most people's understanding of the subject is.
First of all, I'd like to state that evolution is a beautiful theory, although I shall mostly talk about sub-aspect of it known as "the theory of common descent". This part of evolution is most opposed to intelligent design. What we can get from the theory of common descent is perhaps best illustrated by the taxonomy of races -- the tree of life as we often put it.
Here's some experimental facts that support common descent:
we share the same DNA system with all the other animals
It is very likely that the DNA system did not evolve all by itself but we have probably just lost the track of what-was-before-DNA because DNA-based life probably wiped it out altogether.
we share many of the important molecules that power cellular biology, in particular those involved in the citric-acid cycle
Billions and billions of similarly-behaving molecules exist, but nearly everyone has exactly the same molecules involved in the same way, because they were inherited.
we have vestigal traits that were properties of our ancestors in our branch of taxonomy (for instance, humans have vestigal tails and sometimes are born with such) but we do not get vestigal traits that are property of other branches of the taxonomy
This isn't because the tree of life is built just based on vestigal traits, although that is one piece used when humans have puzzled out the speciation. Taxonomy is supported also by knowledge of geography (geographical separation is a requirement for speciation), direct genetic comparisions of DNA samples, and external comparisions of features in animals (although sometimes evolution developes same features over and over again in taxonomically only distantly related animals). These are all mostly together in harmony, although there's a case here and there that biologists still argue over. The overall picture is however very clear.
It's sometimes said that the theory of common descent is not a scientific theory because it's not testable, but that is quite false. The theory of a common descent not only has been observed in laboratory conditions with nothing weirder required but one molecule that is capable of producing copies of itself but it also makes a lot of predictions precisely because it imposes the taxonomy of species. The taxonomy itself is evidence for common descent: every living species must have an ancestor and the genetic material evolves only through mutation and adaptation that is very limited with theory of common descent.
The reason the theory of common descent is a scientific theory is because we can reason with it. Consider, though, that we replace this with intelligent design. Suppose for instance that we would observe that all living things are based on the DNA system. The answer with theory of common descent is that, well, it's that way because every living thing evolved from a single DNA-containing cell so everyone got the same system. However, with an intelligent designer theory humanity is limited to guessing something like, "well, DNA system is the same because maybe the designer was lazy, why would he invent more than one system, when one is enough". You see, it's not quite an explanation at all.
Science is an inquiry about the world and here we must start guessing why the designer would do this or that, and we have no way to determine whether our answers are correct or not. This is exactly why intelligent design can not be taken as a scientific theory, because it's incompatible with the process that advances the state of human understanding.
Taxonomy is the iron claw which will squeeze the life out of theories that do not contain common descent, because the case for taxonomy is extremely strong and leaves no room at all for any species appearing out of nowhere. If you did find one such species, the taxonomy and thus common descent could come crumbling down. But for this, you need experimental evidence that will contradict common descent first.
What about the origin of life?
The answer to that is that it's not very interesting from the viewpoint of common descent. The DNA system is capable of storing incredible amounts of complexity, so all that the theory of common descent needs is that there is a system capable of storing complexity and acting on it and producing copies of itself. So the actual origin of life is not very interesting. The reason this is interesting is because people care about it for philosophical reasons.
The most likely answer, according to my understanding, is that we had some simple molecules combining spontaneously in some energetic soup until something like one RNA molecule was built that then somehow managed to make a primitive living until it seeded the DNA system, at which point the new system proved to be wildly successful and wiped out all the competition.
Now that humanity is approaching the digital life revolution, being able to store and modify the contents of individual DNA, it's rather likely that we take the reins over from blind evolution and direct our own development. It could take a hundred years or two, but it will probably happen.
One other possibility, widely publicized in sci-fi literature is the birth of machine intelligence (and the possible superceding of humans by machines). So in a way, I'd advice to view the current system as a stopgap measure in evolution, which aims towards increasing complexity: it doesn't really matter how it started but where it's going.
If we can create a new species of self-replicating machines, then we in fact do succeed in intelligent design. However, since life as we know it has originated from very primitive stages on Earth, it's probably most likely that it was created here by random processes producing small self-replicating organic molecules. Somehow, from here to there, it must have taken off. We have not, to my knowledge, ever managed to reproduce these steps in laboratory.
What about the capability of evolution to produce complex designs
Considered as given by natural selection. If a change, either mutation or regular in-species variation, improves an organism's chance of reproduction (and/or its ability to nurture offspring, if it nurtures them), then the stage is set for producing more complicated designs, provided that the new design represents an improvement so that it becomes more prevalent in population.