Not a particularly original poll, but something I've been thinking about it a lot again after reading a great post by DSatz in another threat (to very heavily summarize, if I may, he said "don't be afraid to EQ to take out unwanted sound").
To me, the biggest concern about doing it is that what may show up as a "flaw" in one playback system may not on another - particularly with regard to bass, which, if you have a giant subwoofer for instance, may sound perfectly fine on that system, but terrible on headphones (especially small headphones). And of course, all listeners are different, too.
It is an issue I grapple with as I post many of my recordings on a public site, and for me, a lot of the enjoyment comes from knowing other fans are enjoying my recordings. To some degree, that means satisfying "average" taste as well as my own. But of course, if you go too far in that direction, you end up with the heavily compressed, heavily treble-boosted sound that people now take to be "normal."
I'm also curious, when people do EQ, how heavily they are really doing it. Personally, I try to really only roll off low frequencies and/or boost high frequencies by a few dB (+/- 3) unless something is really off. I only play with the presence and midrange to compensate for more unusual flaws, and I've only done that maybe a few times.
I suppose I could have broadened this to discuss compression and other techniques that are common to studio recordings today... feel free to mention how heavily you, for example, use compression to boost "loudness" in the comments.