I haven't read Molyneux's book on the subject, nor did I watch the above video. I read your post and only wished to comment that, from what I've seen, the only working objective moral system is Divine Command theory.
The following is only based on what I read in this post, so forgive any ignorance: I don't know what Molyneux means by "preferable." Something cannot be objectively preferable, unless established by an objective moral code (but that would be begging the question). Is Molyneux settling for a subjective moral code?
I appreciate Molyneux's political videos, being a libertarian, but I must say that his book "Against the Gods?" contained the worst atheistic arguments I've ever heard—ever. One, I think, was something like, 'gods cannot be eternal because they would die of old age.' Seriously. I expect his atheist moral system wouldn't be much better.
He prefaces his book with that exact kind of argumentation. It bothers me a bit because he is so on point with many other things.
Regarding "preferrable," Molyneaux's idea is demonstrated in the example he gives about two men locked in a room. If they both have an ipod, would it be most universally prefferable (he distinguishes between prefferable and preferred) that they each refrain from stealing the others Ipod. It cannot be a universal rule that one steal from the other, since one would be without and the other would have ill-gotten gains. It cannot be universally preferable that they both steal from one another, since no one will have gained and both would have lost. Thus is can only be universally preferred that they each retain what is theirs, and only get when it is agreed to exchange between the two of them.
Concerning your sentence: "It cannot be a universal rule that one steal from the other, since one would be without and the other would have ill-gotten gains." If this is what Molyneux is saying, he can't know what he's talking about. This sentence assumes that it is bad for one to be without and for the other to have ill-gotten gains. He can't assume morality to prove a moral source theory. See what I'm saying? The gains are only "ill-gotten" if we accept a morality in the first place, in which case we don't need his theory at all.
Furthermore, from what you write, it sounds like his theory is nothing more than contractarianism, which fails because it's not universal. That is, if somebody could get away with an immoral act without bad consequences to themselves, in Molyneux's world it would be fine to do it without moral consequence.
I'm sticking with God.
I wouldn't say that he doesnt know what he is talking about, but that this is the best kind of argument you will get without God being the arbiter of good. From his worldview, UPB is just the distilled result of human behavior and applied most consistently. Its not that I necessarily disagree with him, as I agree with his analogy, but that this is simply the result of God's Nature being reflected in what He has made. Molyneaux's approach is to take the result and work backward instead of looking at root causes and working forward.
I would say, rather than Molyneux simply looking at things the wrong way, that his system doesn't in fact work at all. Sure, it may work in most situations, but consider the following:
It is not preferable for slave-owners to free their slaves. It would be preferable for the slaves, but not for the slave owners, because they would have no advantage in freeing the slaves. The question that morality is supposed to answer is: what motive do the slave owners have for releasing the slaves?
In Divine Command Theory: "Thus saith the Lord," and you better listen to him!
In UPB: None.
The point is that a moral source must care about every being in order to be compatible with what our consciences call morality.
In UPB, the moral source, "that which is preferable," is not the same for every individual, and allows people to take advantage if there are no bad (earthly) consequences.
Morality is a question of consequences or judgement. If not, then a person might say, "Why should I bother being moral?" The reason Divine Command Theory alone works is that it is the only moral theory in which all sinful acts are judged. All naturalistic moral theories leave many times when that which our consciences say is immoral does not have negative consequences for the perpetrator—and are, therefore, in these theories, not within the bounds of what characterizes that which is "immoral."
I ought to write an article on the Moral Argument soon.
"It is not preferable for slave-owners to free their slaves. It would be preferable for the slaves, but not for the slave owners, because they would have no advantage in freeing the slaves. The question that morality is supposed to answer is: what motive do the slave owners have for releasing the slaves?"
This wouldn't actually be the case under his system. The nomenclature is important. Universally preferable, not simply preferred. Basically, what would be the most consistent with itself, and for all parties involved. The problem, as has been noted, is that it lacks an authoritative support.
You should read his book UPB. Its free on youtube to listen to