mol, it really isn’t. Read the wiki and the C++ FAQ I linked.
It’s true that usually byte=8 bits, but it isn’t/wasn’t always the case. Historically there were architectures with 9-bit, 10-bit, 12-bit and 36-bit bytes (CDC, PDP-8, PDP-10).
We’ve pretty much settled on x86 inspired architectures now, which are all 8-bit or multiples thereof. So a byte is almost always 8 bits. But it’s not correct to define a byte as being 8 bits.
Also in comms “a byte” may have extra start, stop and parity bits on it, as well as the 7 or 8-bits of data. Hence why they correctly use “octet” to refer to a group of 8-bits in comms specs.