Oh me too, me too--but not soap, stinkbugs. I don't know if you've had the distinctive pleasure of eating a stinkbug by accident (I spent time living in a sufficiently rural area to have blackberry bushes and, well, the stinkbugs liked them too...), but it tastes almost exactly like cilantro. Soap is only kinda the same. At least to me.

Who knows if we have the same taste receptors, though.

Actually, 23andMe could probably tell us. And a fMRI scanner could probably shed some light on whether "like soap" or "like stinkbugs" as a descriptor was supported by our neural activity--assuming we were each willing to sit in a fMRI machine while chomping on cilantro, stinkbugs, and soap. And people who didn't think it was icky would presumably show different patterns. Good times, eh?

(Maybe fMRI wouldn't be precise enough. Maybe we'd need electrode arrays. The point, though is that there is some neural correlate of our sensory experiences, and that in principle could be measured directly.)

But now the validity of my opinion is not quite so assured. Because maybe I'm actually misrepresenting my own qualia by declaring it to be "like stinkbugs" when actually my declaration is more due to my shoddy linguistic abilities when it comes to describing tastes than the actual similarity of experience.

(I, unprovably, declare that this is not the case for me--as I child I called cilantro "stinkbug weed" precisely because the first time I ate it I thought I had eaten a stinkbug along with it. Don't ask why there were so many stinkbugs around, or why I didn't check more carefully!)

So while I agree that some opinions are worthless and you're quite right that "This painting is good" is making an objective claim that may be subject to refutation or at least be of more weight when stated by an expert than a layperson, I think that not even personal preferences are beyond question. Indeed, psychologists have coined the term "revealed preference" precisely for those cases where statements and actions disagree. In these cases we can't always know which (if either) is the "true preference", but it certainly is the case that they can't both be accurate statements.