Is it just me, or is non-consensual sex obviously a bad thing?

"Obviously bad" isn't a utilitarian justification.

Banning Dalits from going within 96 feet of Namboothiris has much more harm done to Dalits than Namboothiris' feelings of ritual pollution. This isn't the case with non-consensual sex.

To play the Devil's advocate:

1. I expect you seriously underestimate the strength of Namboothiris' feelings. To us it seems like pure religious madness, moreover we feel outrage at the extreme inequality existing because of ancient caste prejudices, so we tend to sympathise with the Untouchables and regard the traditional Brahmin rights as unjust. But it doesn't seem that way from the Brahmin perspective.
2. Some of the unpleasantness connected with non-consensual sex is probably status related - being raped makes one lose a lot of status and we tend to avoid status loss. I wonder how much less serious problem would rape become in a society where the negative status effects were removed. We find it acceptable to solve the caste problem by rebuilding the society and changing the people's values - even when many people are objecting; why not attempt the same approach to rape?

(Disclaimer: I think that caste society is unjust and I don't actually wish to change our society to be more rape-tolerant. But I am no utilitarian. This comment is a warning against creating fake utilitarian explanations of moral judgements made on non-utilitarian grounds.)

I haven't been suggesting using (A or B) as a name for (not ((not A) and (not B))) in constructive logic where they aren't equivalent. Rather, I have been suggesting using classical logic (where the above sentences are equivalent) with a constructivist interpretation, i.e. not making difference between "true" and "theorem". But since it is possible for (A or B) to be a theorem and simultaneously for both A and B to be non-theorems, logical "or" would not have the same interpretation, namely it wouldn't match the common language "or" (for when we say "A or B is true", we mean that indeed one of them must be true).

As I understand the responses most people think the main point of Newcomb's problem is that you rationally should cooperate given the 1000000 / 1000 payoff matrix.

I am no expert on Newomb's problem history, but I think it was specifically constructed as a counter-example to the common-sensical decision-theoretic principle that one should treat past events as independent of the decisions being made now. That's as well how it is most commonly interpreted on LW, although the concept of a near-omniscient predictor "Omega" is employed in wide range of different thought experiments here and it's possible that your objection can be relevant to some of them.

I am not sure whether it makes sense to call one-boxing cooperation. Newcomb isn't Prisoner's dilemma, at least in the original form.

OK, I understand now that your point was that one can in principle avoid being predicted. But to put it as an argument proving irrelevance or incoherence of the Newcomb's problem (not entirely sure that I understand correctly what you meant by "dissolve", though) is very confusing and prone to misinterpretation. Newcomb's problem doesn't rely on existence of predictors who can predict any agent in any situation. It relies on existence of rational agents that can be predicted at least in certain situations including the scenario with boxes.

I still don't understand why would you be so much surprised if you saw Omega doing the trick hundred times, assuming no stage magic. Do you find it so improbable that out of the hundred people Omega has questioned not a single one had a quantum coin by him and a desire to toss it on the occasion? Even game-theoretical experiment volunteers usually don't carry quantum widgets.

Tell that to the hypothetical obscurantist.

Edit: I find it mildly annoying when, answering a comment or post, people point out obvious things whose relevance to the comment / post is dubious without further explanation. If you think that the non-equivalence of the mentioned beliefs somehow inplies the impossibility to extrapolate obscurantist values, please elaborate. If you just thought that I might have commited a sloppy inference and it would be cool to correct me on it, please don't do that. It (1) derails the discussion to issues of uninteresting nitpickery and (2) motivates the commenters to clutter their comments with disclaimers in order to avoid being suspected of sloppy reasoning.

Couldn't this be said about any inductive method, at least in cases when the method works?

There are obscurantists who wear their obscurantism as attire, proudly claiming that it is impossible to know whether God exists. It can be said, perhaps, that such an obscurantist has a preference for not knowing the answer to the question, for never storing a belief of "the God does (not) exist" in his brain. But still all the obscurantist's decisions are the same as if he believed that there is no God - the obscurantist belief bears no influence on other preferences. In such a case, you may well argue that the extrapolated volition of the obscurantist is to act as if he knew the answer and therefore the obscurantist beliefs are shattered. But this is also true for his non-extrapolated volition. If the non-extrapolated volition already ignores the obscurantist belief and can coexist with it, why is this possibility excluded for the extrapolated volition? Because of the "coherent" part? Does coherence of volition require that one is not mistaken about one's actual desires? (This is a honest question; I think that "volition" refers to the set of desires, which is to be made coherent by extrapolation in case of CEV, but that it doesn't refer to beliefs about the desires. But I haven't been interested in CEV that much and may be mistaken about this.)

The more interesting case is an obscurantist who holds obscurantism as a worldview with real consequences. Talking about things that are plausible (I am not sure whether this kind of obscurantists exist in non-negligible numbers), imagine a woman who holds that the efficacy of homoeopathics can never be established with any reasonable certainty. Now she may get cancer and have two possibilities for treatment: a conventional, with 10% chance of success, and a homoeopathic, with 0.1% chance (equal to that of a spontaneous remission). But, in accordance with her obscurantism, she believes that assigning anything except 50% for homoeopathy working would mean that we know the answer here, and since we can't know, homoeopathy has indeed success chance of 50%.

Acting on these beliefs, she decides for the homoeopathic treatment. One of her desires is to survive, which leads to choosing the conventional treatment upon extrapolation, thus creating conflict with the actual decision. But isn't it plausible that her another desire, namely to ever decide as if the chance of homoeopathy working were 50%, is enough strong to survive the extrapolation and take precedence upon the desire to survive? People have died for their beliefs many times.