Gary Francione argues that non-human animals have the interest to live, that we should take their interest of living into account, and that giving more importance to our interest of eating them is speciecist. I do not think it is speciesist, because if I eat meat I don't care if it belongs to a non-human animal or to a human, I'll eat it just as well regardless of the species (besides worrying about kuru or prion disease). Putting aside the need to change my diet in order to be consistently utilitarian, this brings me to the scenario I'll describe.

Yes, let's say non-human animals have an interest in living, one to be taken into account from a preference utilitarianism perspective. However, they do not understand death to the larger extent humans realize it. I doubt the death of a non-human animal causes half the amount of consequences in their corresponding society than the death of a human causes in human society. E.g. the mother who lost his son in a car accident could advocate and promote automobile safety, saving other people's live's as a result.

Now imagine a world with two groups of humans, Group 1 and Group 2. Group 1 is oblivious of Group 2, and it is controlled by Group 2. Group 2 is extremely advanced technology-wise, enough to satisfy Arthur Clarke's third law: their advanced technology is indistinguishable from magic. Group 1 lives in an underground facility à la Brave New World / Michael Bay's The Island. They live happily in what they think is all there is in the universe, the underground facility which looks like some sort of island or beautiful place of which they cannot escape nor want to. They live happy satisfying lives, until they reach the age of 30, when they die peacefully in their bed. They think this is normal, they are unable to avoid it, they see it as the inevitable cycle of life. Humans of Group 2 have arranged the deaths of humans of Group 1 at the age of 30 through technology, and use those bodies from Group 1 as part of their diet. Humans of Group 1 will never ever have the chance to know and understand this. People from Group 2 feed from dead people from Group 1, and they have found that indeed this is the scenario that causes the most net happiness. They make sure people in Group 1 are very satisfied, and they are. All people in Group 2 are okay with this. Despite their different destinies, humans of both groups are of the same species: homo sapiens sapiens, and no harm is done by eating humans since medical science is very advanced.

In this scenario, would you agree with what Group 2 is doing, knowing that it causes most net happiness?

I would definitely agree with Group 2. I would also agree with killing someone painlessly if their life isn't worth living, if they're using resources that could be allocated more efficiently, if they committed a horribly crime and killing them would serve as an effective warning to others, etc. I don't see why food should be any different.

Frey's book on Interests and Rights attributes the interests paradigm to Leonard Nelson's A system of Ethics from 1956. I haven't read that book, but I might be able to get it on Thursday. Some have accepted the interests challenge, and some have not. Some who accepted it argued in favor of animal rights, and some have not.

My first question to myself is whether I accept that statement as something interesting to ponder or if I accept it as a "law" of ethics. I take it as interesting but I see it as yet another attempt to fabricate a physics-like law of ethics without having a rational basis for doing so.

I am skeptical of any process of inferring ethical laws from observation unless it comes exceedingly close to complete enumeration of all possible cases that might be affected by the law -- and I am skeptical of human ability to do this. The process of observing physical phenomena and inferring laws has developed over the centuries and has strong support from many domains of observation and has strong support for its intersubjective reliability (many people looking at the same thing and all of them giving the same answer). Moral philosophers have been trying to infer ethical laws for millennia and produced no agreement on anything. The most defensible inference from that track record is that human beings, as a species, exhibit an exceedingly strong tendency to make false inference about moral laws and then maintain strong overconfidence in those inferences: an illusion of insight.

This is how I perceive the interests theory and its origins: there is a metaphysical entity that exists in reality and says that entities get rights if and only if they have interests. Then another metaphysical entity somewhere in the universe defines what "interests" are. Then some mysterious attribute of humans that has lain dormant from the beginning of time until 1956 emerged through revelation and now informs philosophers left and right with the content of this metaphysical knowledge, and this information is not thrust into our minds until was ask ourselves the question.

Singer's utilitarianism does not rely on a conception of rights, so it does not rely on a definition of interests and does not rely on the human ability to divine this information or the human ability to objectively identify the living organisms that possess this quality.

Utilitarianism might be messier than acting as if I had metaphysical insight into a clear policy that tells me what right and wrong are, but my desire to have perfect answers that conform to a newly-discovered natural law is not sufficient to overpower my desire for an accurate understanding of the universe around me. My desire to understand things accurately is instrumental to my larger goal of pursuing my best long-term well-being/happiness (because accurate information is essential to long-term choice of actions).

As for your vignette, it would be more humane than the current factory farms, but it disregards the well-being of others, and I would be happier if people could stop killing others. If humans develop the technology and resources to build such a system, maybe they should adjust it to help people voluntarily reduce their desire to eat other humans and instead trick themselves into really enjoying rice and barley mush.

Firstly, it's worth noticing that just because the Brave New World is portrayed as utilitarian, doesn't mean we can take it to be so. See David Pearce's critique here. But taking the question as intended, if we could get humans to live happy lives, would we eat them. It seems to me that the answer is yes but there is a long list of disclaimers.
Firstly, the taboo surrounding cannibalism is here to stay.
Secondly, so it should! There's no forseeable circumstance in which it would be useful for humans to be more tolerant of eating each other. This is simply unlikely to contribute to the cooperation between people in a flourishing society.
Thirdly, the question distracts from the wrongness of omnivorism. If we conclude that eating other beings is not intrinsically wrong, then we may better tolerate the eating of animals. The practical question is not whether we should eat happy beings. The question is whether we should eat sad beings. And the answer is a decisive no. As David Pearce has also said, any utilitarian who eats animal products is either weak-minded or weak-willed.

So the utilitarian position is that eating other beings is not intrinsically wrong. But nor in the utilitarian position is anything wrong, bar suffering. So, in response to the question 'is eating beings wrong?', my initial answer was 'no, not intrinsically'. But an equally valid answer is 'yes, practically. Just as are torturing and killing, the eating of beings is wrong'. It ought not be done in any forseeable circumstance. It ought to be prevented. It ought to be condemned.

The thing is, the isolated incident of killing and eating a living being is wrong in the sense that it stops him from continuing to be happy if he was so. But what if the fact that the living being was happy while living was precisely because he was going to be killed at a later date? I.e. the animal breeder breeds animals in joyful conditions (I am aware this is not what actually happens, but bear with me in the hypothetical) so that he'll later kill them to have them for food. He would be generating happiness that otherwise wouldn't exist, since the animals he's caring for wouldn't be there to begin with. In this sense breeding animals for food would be good.

The only utilitarian convincing objection I see to this ideal farming situation is that the time and work dedicated to it could be better spent doing something a lot more effective in generating happiness. Which equates the moral wrongness of farming animals in ideal conditions to the moral wrongness of producing the videogame Duke Nukem Forever or Spider-Man The Musical: lots of millions of dollars spend for little enjoyment produced.

The worst thing I see in eating animals is not the act of killing them for it, but all the suffering it entails within the farming industry.

I found the book by Nelson (A System of Ethics, 1956 -- appears to be a translation of the 2nd German addition from about 1949). He allegedly invented the idea that interests are the basis for rights and that animals are included in this. The book was filled with synthetic propositions in an attempt to build a logic of ethics. I wouldn't have bothered reading it if others hadn't referred to it. It also doesn't say much about your scenario, although he did invent an imperative that says "respect personal dignity!" (p. 98). So if you want to listen to that guy, you shouldn't deceive humans as part of your plan to eat them.

a few other quotes:

"On the basis of the principle of personal dignity each person is entitled, by virtue of his interests, to restrict the will of ohters" (p. 98). [this doesn't seem to coincide with the dignity imperative]

"every person has a claim to respect for his interests" (p. 99)

"Any being that can feel pleasure or pain is therefore a subject of rights, and has dignity in the sense defined above." (p. 100).