They are not special in a regex.

That is true for Perl, but some regex engines (GNU) use \< and \> as word boundary anchors, where \< is beginning of word and \> is end of word, whereas Perl only has \b. The argument against \b is that it is easily confused with a backspace character, and that it does not differentiate between start and end of word.

So, using \< and \> could be doubly confusing to anyone coming from other RE regimes. Matching a string starting with a word-ending?