Harris sells us utilitarianism and utilitarianism alone. His views would be much more palatable if he distinguished between the moral theory he is espousing and his argument that science can help us interpret/apply the theory.
I don't think it is particularly well founded for a reason that hasn't been mentioned yet:
http://en.wikipedia.org/wiki/Deontological_ethics
http://en.wikipedia.org/wiki/Consequentialism
I believe that deontological ethics - that is, coming to decide right and wrong actions - is superior to consequentialist ethics, which is to judge actions by their expected results.
Many people would disagree with me - a lot of people are consequentialists. But consequentialism usually fails, once you start saying "The ends justify the means" you get into bad places really fast.
Historically speaking, consequentialism has produced lots of problems and not much success.
It's also possible to be consequentialist and still refuse to due certain things on principle, even if it seems like the right thing to do, because you know you are prone to be wrong about that kind of thing.
http://lesswrong.com/lw/uv/ends_dont_justify_means_among_hum...
This is a good point, but it appears like the process of attempting to apply consequentialist thinking almost always leads to bad places. Consider, for example, Prohibition or the War on Drugs. These were both really well-intentioned in some ways - they were about having a cleaner, more sober, more healthy society, with less poverty and more prosperity, and less addiction. But the end result is creating an underground drug/alcohol trade, giving rise to organized crime and gangs, and having people now consume less safe homebrewed versions of alcohol and drugs.
I think most if not all well-intentioned consequentialist thinking goes to similar places. There were will always be seductively alluring "Works in theory/destroys the world" type things (separate but equal segregation, communism, prohibition, price controls) - and many of those are in violation of basic deontological ethics of giving people choice, letting them live their lives, and so on. I think the vast majority of drastically bad ideas can be avoided with straightforward, rule based live and let live type ethics, which is what I personally favor after searching around.
"Ends justify the means" has historically always led to bad places - and I think it always will. "Ends justify the means" works in theory, but has consistently historically failed in the real world.
Edit: Just read the Less Wrong article - quite good and I more or less agree with Eliezer there.
ADDED: Thinking about it further, the only people who use deontological ethics any more that I can think of are religious people and Marxists, those who have a transcendental source for their required actions.
Don't get caught up on exactly what words he used, but rather watch the whole talk and try to get his intent, which I think is well founded.
He is, I think, fairly clear in saying that science does not tell us what specifically is morally "good" or "bad," but rather that science can tell us what human activities foster and correlate with human prosperity and happiness. I'm sure most everyone here will agree that this is not an outrageous claim. Since moral questions usually deal with issues of prosperity and happiness, he concludes that we shouldn't be afraid to bring up science in such discussions.