|Keep It Simple, Stupid|
Programmers will be necessary until someone solves the halting problem. The automatic programming problem (writing a program that writes programs for you) has been proven NP-complete.
Even then, programmers will be necessary, because what programmers really do is translate human assumptions into symbols that a computer can manipulate. OOP, declarative programming, functional programming, descriptive programming.. they're all just frameworks on which to arrange our thoughts. Perl, C, C++, Java, Lisp.. they're just collections of symbols and primitive operations that we hang on a particular framework.
The really hard part of programming is learning to think without contradicting oneself. When you tell a computer to contradict itself, it can extrapolate that contradiction into a database full of gibberish right smart quick. Humans, OTOH, are so good at dealing with ambiguity that we can hold self-contradictory opinions without ever noticing.
Human thought is by nature fuzzy and ill-defined. We can identify a bear well enough to run away from it, while we're still far enough away that running will work. We don't care about its exact weight, or how many hairs are in its pelt, because finding out has never been a major survival trait.
Computers, by contrast, are excruciatingly precise. the C code:
will fail as often as not, due to microscopic differences in the interpreted values of those two variables.
Programmers take ideas from the fuzzy, human, "oh crap, a bear!" realm, and polish them until they work in the meticulous, "can't tell if 1 really equals 1" computer realm.
That's not easy.
You can always find people who'll say that computers should be able to write programs that whatever we tell them we want, and in a sense, that's true. The problem is that most people don't know what they want well enough to explain it to another human, let alone a computer.