Introspection is not particularly trustworthy. If you consciously (as opposed to acting on auto-pilot) decide to move to "an even better state", you have in fact evaluated your current conscious state and concluded that it is not the one you want to be in, i.e. that something (at least the fact that you'd rather want to be in some other state) bothers you about it. And that -- wanting to get out of your current conscious state (or not), or changing some aspect about it -- is what constitutes suffering, or whatever (some) negative utilitarians consider to be morally relevant.
If you accept this (Buddhist) axiology, it becomes a conceptual truth that conscious welfare-tradeoffs always include counterfactual suffering. And there are good reasons to accept this view. For instance, it implies that there is nothing intrinsically bad about pain asymbolia, which makes sense because the alternative implies having to go to great lengths to help people who assert that they are not bothered by the "pain" at all. Additionally, this view on suffering isn't vulnerable to an inverted qualia thought experiment applied to crosswiring pleasure and pain, where all the behavioral dispositions are left intact. Those who think there is an "intrinsic valence" to hedonic qualia, regardless of behavioral dispositions and other subjective attitudes, are faced with the inconvenient conclusion that you couldn't reliably tell whether you're undergoing agony or ecstacy.
Introspection is not particularly trustworthy.
My point was that your favorite theory cannot make sense of what people take themselves to be doing in situations such as those discussed above. You may argue that we shouldn't trust these people because introspection is not trustworthy, but then you'd be effectively biting the bullet.
...If you consciously (as opposed to acting on auto-pilot) decide to move to "an even better state", you have in fact evaluated your current conscious state and concluded that it is not the one you want to be in, i.e.
I've always been more of a theoretician, but it's important to try one's hand at practical problems from time to time. In that vein, I've decided to try three simultaneous experiments on major Less Wrong themes. I will aim to acquire something to protect, I will practice training a seed intelligence, and I will become more familiar with many consequences of evolutionary psychology.
In the spirit of efficiency I'll combine all these experiments into one:
She's never seen Star Wars or Doctor Who.
She's never seen David Attenborough or read J. L. Borges.
She's never had a philosophical debate.
She's never been skiing.
Never had sex, never been hugged or even been licked by a dog!
She has so much to look forwards to...
(Though she'll be very boring for several months yet!)