They always put the small fonted:
1KB = 1000B; 1MB=1000KB; 1GB=1000MB; 1TB=1000GB.
Whereas the real amounts are:
1KB = 1024B; 1MB=1024KB; 1GB=1024MB; 1TB=1024GB.
I'm afraid the "real" amounts, as in standardized (SI) amounts are actually the ones above, while the ones below should be (by IEC standard) KiB,MiB,GiB,TiB. By another standard (JEDEC's) KB (as opposed to kB),MB,GB (but not any units above those (e.g. TB)) may also be used as the equivalents of KiB,MiB,GiB. That was introduced due to common (false) usage of KB,MB,GB back then (JEDEC standardized it in 2002).
I was convinced myself before reading some wiki articles after first reading this topic a few days ago, as my memory apparently failed me, that K(B),M(B),G(B),T(B),... were the binary (exponents of 2) prefixes and that the Ki(B),Mi(B),Gi(B),... were the decimal prefixes.
This should clear this up. It also describes the different usages in networking, storage, software products and elsewhere.