I'd say UTF8 BOM isn't deprecated, but it might be discouraged in certain filetypes, but then in exchange for other ways of telling the same (such as in xml and html)..
The reason is that without the BOM (or similar things such as html charset and xml encoding etc) there is no way to determine 100% accurate whether the file
uses utf8 or an ansi codepage (the latter is often still the case).
Do with them? as in provided by the OS?
Maybe not, but then for most people the OS just sits there, the actual programs people tend to use is more than those. and those often do support utf8 (apparently more often
than not nowadays)