The two cases are not easily comparable, because the current states of evidence are very different. Also, saying the correct GUT might not yet be conceived is like saying that the correct exotic dark matter particle might not yet be conceived. GUTs and exotic dark matter are both classes of theories; saying we don't know the right one in the class isn't the same as saying we should look in another class.

In general, proton decay is a prediction of GUTs because they do away with conservation of baryon number. (You're right that Pati-Salam has this implicitly, though not explicitly.) It may just be that proton decay is very very suppressed, so the experimental half-life bounds of ~10^34 years would be acceptable. Some newer GUTs (utilizing supersymmetry, for example) make predictions that fit the experimental half-life bound. Supersymmetry corrections make the electromagnetic, weak, and strong forces all unify at a single point. Supersymmetry also predicts a new stable particle (the lightest supersymmetric particle) which could be a dark matter candidate.

I'm not sold on all of that, though I admit it has some appeal. No direct evidence yet exists. If supersymmetry is correct, we should find evidence of it at the LHC. Because of the dearth of evidence, however, this is an area where current hypotheses may just be off in the wrong direction. If that's the case, whatever actual evidence we get from the LHC should serve as a course corrector.