> Add to that the automatic CAPTCHA-hellban you get if you use "site:" in anything more than a tiny amount

Pretty much any advanced operators seem to do it for me, notably "intitle:" and "inurl:". I'd wager that there are a lot of automated searches using these to look for exposed admin interfaces, but I find them extremely useful for filtering out the crap that clogs up results when a ton of news sites all regurgitate the same viral press release or wire article.