Five letters, five code units
In UTF-16, `hello` is `0x00680065006c006c006f`: each letter occupies one 16-bit code unit. A code unit is storage for an encoding, not necessarily a whole Unicode codepoint. section 2.1.2
UTF-16 uses one code unit for codepoints through U+FFFF and two for codepoints above that range. The two-unit sequence is a surrogate pair; its units must be combined to recover the codepoint. For example, `0xD801` and `0xDC37` combine to give `0x10437`. section 2.1.2
When one symbol becomes two matches
Without the `u` flag, a JavaScript regular expression treats the surrogate pair in `😀ABC` as two characters; with `u`, matching is Unicode-aware. Do not assume a code-unit boundary is a codepoint boundary. section 2.1.2

