It's worth distinguishing 'happiness' (positively valenced experience) from preferred experience. For example, an ascetic monk may wish to avoid strongly valenced experiences, or a self-loathing individual might even want to feel miserable (i.e. have negatively valenced experiences). More interestingly, we might feel attached to our current tastes (hedonic likings), and so be averse to the idea of inducing satisfactions of other tastes.
Imagine, for example, an aesthete who loves opera and fine wine, and utterly detests Britney Spears and cheap bubbly. Suppose a mad scientist offered to rewire their tastes and have them enjoy listening to Britney for hours on end. It's entirely conceivable that the aesthete would find this prospect abhorrent. While acknowledging that he would - in the proposed scenario - be deluded into experiencing great happiness, from his actual perspective the aesthete strongly prefers not to undergo this experience. He prefers not to enjoy the experience of listening to crap.
Is this preference unreasonable? I think that Derek Parfit may be committed to saying so. Parfit holds that hedonic likings and dislikings (unlike desires) are not responsive to reasons. They are simply brute psychological facts. We can have reason to want to like fine wine, if it has a greater potential for intense liking/enjoyment than does cheap bubbly. But there is no reason to like fine wine (any more than there are reasons to bleed), if 'liking' is a brute, reasons-insensitive state: something that happens to us, rather than something we do.
Anyway, if Parfit is right that there are no reasons for liking aesthetically superior objects more, then it's hard to see why we should care about the objects, as opposed to the intensities, of our likings. That is, it looks like the only reason to want to like fine wine more is insofar as it would yield a more intense liking. This might seem to devalue aesthetic evaluation. Beethoven's symphonies no longer merit appreciation. They're just cheesecake. Really, really enjoyable cheesecake. And if you could find Britney just as enjoyable, then that would be just as good.
I wonder about applying this to the moral case. It seems intrinsically bad to derive pleasure from others' suffering (even if you reflectively prefer that others not suffer, and would never act to cause suffering, or anything like that). If a mad scientist offered to rewire your tastes to make you enjoy child porn, say, that would seem very undesirable indeed (and again, I think, not just on instrumental grounds -- though that would surely be part of it). Are these intuitions mistaken?