Should be easy enough for networks to detect and remove these, by identifying content where character ranges in words routinely fall outside the charsets of lan

Should be easy enough for networks to detect and remove these, by identifying content where character ranges in words routinely fall outside the charsets of languages. That, or some sort of fuzzy CV hashing, which is cool, but more intensive. That would also mitigate null length and invisible modifiers.