Square-bracket indexing of code units within a string isn't a "modern alternative" tocharAt, both have been in JS since the beginning. The only difference is square-bracket indexing returnsundefined instead of"" for out-of-bounds indexes.

NeithercharAt nor square-bracket indexing are truly modern alternatives, because neither support non-BMP characters (e.g. most emojis):

'💩'[0]// '\uD83D''💩'.charAt(0)// '\uD83D'

You can make a more modern version that works generically on iterables something like this:

functionatIndex(iter,idx){if(idx<0)return[...iter].at(idx)leti=0for(constelementofiter){if(i++===idx)returnelement}}

Usage:

console.log(atIndex('💩 abc',0))// 💩console.log(atIndex('abc 💩',-1))// 💩console.log(atIndex(newIntl.Segmenter('th-TH').segment('แปลให้หน่อยได้ไหม'),4,)?.segment,)// ห้

Strings are iterated by codepoint, rather than code-unit, so emojis and other non-BMP characters are handled correctly.

The last example uses segments from anIntl.Segmenter as the iterable, so each "character" is a grapheme cluster (closer to the concept of what users of relevant languages would consider to be a "character"). A grapheme cluster can consist of one or more Unicode codepoints.