Kibibytes are silly and we should all use them
The kibibyte and mebibyte and their ilk are kind of odd. Introduced in 1999, these units are close to the kilobyte, megabyte, and so on, but they’re related to each other by factors of 210 = 1,024 rather than the factors of 103 = 1,000 used by the more familiar SI prefixes.1 The sloppy habit of using “kilobyte” to mean 1,024 bytes was becoming more and more common, and the kibibyte was a way to preserve this (admittedly useful) unit while distinguishing it from the true kilobyte.
There are downsides. Converting between the binary-based units is more involved than just moving a decimal point, as you can do with the 10-based prefixes. They have unfamiliar names, and the double “b” even sounds a little silly to my ear. The main advantage of the newfangled units is that they’re unambiguous: because of the history here, when someone says “kibibyte” you can be virtually certain that they mean 1,024 bytes.
But when it comes to units of measurement, being unambiguous is the whole point! This advantage by itself is why you should prefer the binary units almost any time you’re talking to other people.2 When you use kibibytes, mebibytes, and so on, no one has to second-guess what you meant.
