That just shows how agile PG is. He added unicode support the minute he saw people request it! :)
Seriously, PG explicitly claims that Arc intentionally doesn't support anything but ASCII (http://www.arclanguage.org/), so that might be why people (including me) believed that to be the case.
Yes, I think Arc intentionally supports only ASCII just to not bother with Unicode issues as of right now.
Anyway, I can't see how Unicode can break in Arc. I'm not a Lisper, but I think you can't extract 1 byte from an Arc string (since it's just a MzScheme string), but 1 char instead. That's a different concept, because in Unicode 1 char can be formed with 1, 2 or more bytes.
Watch out - that's single-portable-implementation thinking. When Paul puts out another release of Arc based on, say, another Scheme implementation, or SBCL, those tricks won't work.