I think 3) is your strongest point, may I try to expand on it?
I wonder, what is Sam's response to utility monsters, small chances of large effects and torture vs. dust specks? In saying that science can answer moral questions by examining the well-being of humans, isn't he making the unspoken assumption that there is a way to combine the diverse "well-being-values" of different humans into one single number by which to order outcomes, and, more importantly, that science can find this method? Then the question remains, how shall science do this? Is this function to be found anywhere in nature? Perhaps in the brains of conscious beings? What if these beings hold different views on what is "fair"?
I simply can't imagine what one would measure to determine what is the "correct" distribution of happiness, although that failure to imagine may be on my part.
Sam would be subject to all the usual objections to utilitarianism, altruism, and moral objectivism available in the existing literature. He has justified not addressing that literature with a glib comment that he was sparing people from boredom. As I said before, he is fundamentally unserious and even dishonest in arguing his case.
Sam Harris is here offering a substantial amount of money to anyone who can show a flaw in the philosophy of 'The Moral Landscape' in 1000 word or less, or at least the best attempt.
http://www.samharris.org/blog/item/the-moral-landscape-challenge1
Up to $20,000 is on offer, although that's only if you change his mind. Whilst we know that this is very difficult, note how few people offer large sums of money for the privelage of being disproven.
In case anyone does win, I will remind you that this site is created and maintained by people who work at MIRI and CFAR, which rely on outside donations, and with whom I am not affiliated.
Note: Is this misplaced in Discussion? I imagine that it could be easily overlooked in an open thread by the sorts of people who would be able to use this information well?