It’s a “joke” because it comes from an era when memory was at a premium and, for better or worse, the English-speaking world was at the forefront of technology.

The fact that English has an alphabet of length just shy of a power of two probably helped spur on technological advancement that would have otherwise quickly been bogged down in trying to represent all the necessary glyphs and squeeze them into available RAM.

… Or ROM for that matter. In the ROM, you’d need bit patterns or vector lists that describe each and every character and that’s necessarily an order of magnitude bigger than what’s needed to store a value per glyph. ROM is an order of magnitude cheaper, but those two orders of magnitude basically cancel out and you have a ROM that costs as much to make as the RAM.

And when you look at ASCII’s contemporary EBCDIC, you’ll realise what a marvel ASCII is by comparison. Things could have been much, much worse.