Part of the problem is that fixing Newton is more than just adjusting the equation; it requires making up a whole new constant, which to some is worse than making up a whole new class of matter. Constants don't have explanations; they're just like saying, "I don't know why it works out this way, it just does." It's better to have a reason that explains why something happens than to throw up your hands and say "let's just add a new constant to make things work out."
A constant is a lot more constrained (existing in a specific context) than a class of matter that can't be observed, which is used as a hand wave to any measurement that is unexpected because not only is it not observed, it's not fully understood/constrained.
That's not terribly convincing argument. Effectively MOND means adding a single free parameter to the model and dark matter means adding a entire continuous distribution to the model.
