Just so that it is not necessary to reconstruct this every time the question of the adequacy of Universally Preferable Behavior arises, I'll just place these comments (originally from Reddit) here for the benefit of posterity:
[To see the problems with the book] you might look at the beginning of the section "UPB: Five Proofs," where he lays out the supposed logical proof from UPB. When he lays it out as a syllogism, the key proposition is that: "Arguing against the validity of universally preferable behaviour demonstrates universally preferable behaviour." Now, that's not the clearest of sentences, but here's the explanation he gives: "if I argue against the proposition that universally preferable behaviour is valid, I have already shown my preference for truth over falsehood – as well as a preference for correcting those who speak falsely. Saying that there is no such thing as universally preferable behaviour is like shouting in someone’s ear that sound does not exist." Now, the power of the argument rises or falls on the assumption that arguing against something always demonstrates the preference for truth and correcting error. And there's nothing self-evident about that claim. The analogy between "universally preferable behavior" and sound doesn't help anything either, since if sound did not exist, no amount of shouting in someone's ears could prove otherwise. The logic of the argument depends on the action demonstrating a particular preference, and always the same preference, but if you had never heard of UPB and had to derive its existence from the fact that I oppose the notion, well... I think there are a variety of explanations which present themselves as at least as likely.
The question of "implicit intent" is a tough one, particularly since Molyneux places intent as such outside the realm of what we can study. So if we are going to make a "universal argument" rest on conclusions we draw about something which we've started off by saying we can't draw any direct conclusions about, then we had better have a strong argument that a given action must demonstrate a constant preference. Market anarchists might compare their assumptions about the dynamics of action in the market with those that Molyneux brings to the table in the UPB argument. Might we not just as easily speculate about the marginal utility of arguing, are about certain kinds of arguments (for or against, in good faith or not, etc.)?
The point about shouting when there is no sound is that it is not logically contradictory, but bizarre, futile and physically impossible. On the other hand, if there is no UPB, one can still argue, can still prefer truth under the given circumstance, can still attempt to correct others, prefer to correct others under a given circumstance, etc. The analogy is pretty misleading, part of the enormous amount of stuff in Molyneux's book which is not particularly to the point, but maintains this barrage of ridicule against any opposing position. And Molyneux does like his ridicule, even when it stretches logic and truth -- so what does that sort of argumentation imply for the UPB?
There is, no doubt, more that could be said, but I think anyone who looks at the book, and particularly the passage in question here, can pretty quickly figure out what they think about the whole issue.