I don't really disagree that all human actions have self-interest as the core motive. I'm just saying that in an interdependent social environment, a phenomenon best described as altruism naturally emerges, either because instincts initially developed for self-serving purposes develop a life of their own, or because the still-existing individual benefits are obscured (i. e. a sort of game-theoretic justification for a middle class person to support welfare).
I'm mostly taking offense with the characterization of such "feelings" as "not real" or of lesser value than the "pure" thought of rational egoism:
1. If we take altruism as a highly developed extension of egoism, it's worth to consider it an invaluable heuristic trained over thousands of generations to intuitively make good decisions. Where a cold analysis may tell you that giving money to a poor person is wasteful, I'd say that the warm fuzzy feeling we get from giving is an encoding of results gained over millennia of experience.
This also touches another point: I'd suspect that most altruistic behaviors are actually net-negative for the individual but, when practiced by large parts of the population are beneficial for everyone. Here, emotions serve as a prehistoric solution to the tragedy of the commons. It's an idée fixe in some circles, most notably libertarians and the Ayn Rand crowd to replace this with a system of contracts or simply force but that disregards (amongst other things) that almost no idea in psychology has more empirical support than the beneficial effects of helping.
2. If we consider altruism as an emergent behavior, a sort of instinctual moral code, it should be revered as one of the highest achievements of humanity on a level with art or science.