I am far closer to your position than sark's, I must admit. Lukeprog provided examples of how we can be wrong about our subjective experience, and commenters reached for more abstract subjective experiences; some of which we can also be wrong about (ie Alicorn's peripheral vision card experiment, the story about blind echolocation believed to be "forehead touch") and some of which are defined self-referentially so that we can't be wrong about it, purely by the virtue of the answer we give being the answer required. The self-referential subjective experiences (such as "I experience myself giving the answer 5 to the question 2+2=?") aren't useful; they're tautological. When it comes to subjective experiences that actually do work, we can be wrong. The only work that tautological subjective experiences do is contradict lukeprog's claim.