Should work with all Unicode that the font your browser uses has glyphs for. You see UTF-8 decoded and rendered, I don't think Arc does anything special with it. When you look at the page source you can see that the “↔” isn't given as an escape sequence or an HTML entity.
Hacker News strips certain characters out. That specific one is not one of the ones that is removed.
I stand corrected. What's the point of doing that?