A byte is a collection of bits - now almost universally standardized on 8 bits because of network protocols and encodings - but older computers the byte wasn’t as important as THE WORD which was the bit width the machine could process at once. We now refer to 128bit instructions but those are more properly an instruction with a 128bit word.

Lisp provides the ability for arbitrarily defined bytes.

so i wasn't totally wrong. It's mainly semantics.

Since i'm here, why do most storage mediums i.e usb, phones sd ostensibly follow this method (1gb 2,4,8!,16,32 etc) and why are they almost always a few hundreds mb off?

A kilobyte can be defined as either 1024 bytes or 1000 bytes. Similarly, a megabyte could be 1024 kilobytes or 1000 kilobytes, and so forth. Really, 1024 is the number that makes more sense, but hard drive manufacturers promoted the “1000 bytes” interpretation so they could advertise their hard drives as having more megabytes.

Some standards organizations tried to “solve” the issue by defining a “kibibyte” as 1024 bytes and so forth, but you only ever see those in abbreviated forms (ie KiB). I personally hate that convention; the words are aesthetically atrocious and were only invented to please pedants (who annoy me) and dishonest marketing people (whom I detest even more).

Isn't this the difference between speed notation and size notation?

Some older machines had different byte lengths. I believe Burroughs had 7-bit bytes. PDP-10s had 36-bit words which could be divided into any byte size that you liked; 7 bits was common, but if you needed them you could get 1-bit bytes, 6 bit (which was used to encode filenames), 8 bits, whatever.