Who, whoami, and “who am i”

Yes it's a joke, included in by the developers of the who command. See the man page for who.

excerpt

If FILE is not specified, use /var/run/utmp. /var/log/wtmp as FILE is common. If ARG1 ARG2 given, -m presumed: 'am i' or 'mom likes' are usual.

This U&L Q&A titled: What is a "non-option argument"? explains some of the terminology from the man page and my answer also covers alternatives to who .. .... commands.

Details

There really isn't anything special about am I or am i. The who command is designed to return the same results for any 2 arguments. Actually it behaves as if you called it with its -m switch.

   -m     only hostname and user associated with stdin

Examples

$ who -m
saml     pts/1        2014-01-06 09:44 (:0)
$ who likes candy
saml     pts/1        2014-01-06 09:44 (:0)
$ who eats cookies
saml     pts/1        2014-01-06 09:44 (:0)
$ who blah blah
saml     pts/1        2014-01-06 09:44 (:0)

Other implementations

If you take a look at The Heirloom Project, you can gain access to an older implementation of who.

The Heirloom Toolchest is a collection of standard Unix utilities.

Highlights are:

• Derived from original Unix material released as Open Source by Caldera and Sun.

The man page that comes with this who in this distribution also has the same "feature", except it's more obvious.

$ groff -Tascii -man who.1 |less
...

SYNOPSIS
       who [-abdHlmpqRrstTu] [utmp_file]
       who -q [-n x] [utmp_file]
       who [am i]
       who [am I]
...
...
       With the two-argument synopsis forms `who am i' and  `who  am  I',  who
       tells who you are logged in as.
...
...

The backticks (``) are command substitution: they are replaced by the result of running the command inside the backticks. Here they run whoami, which prints your username.

The - after su makes su run a login shell: a login shell will read certain environment configuration from scratch, among other things. By default it would just run the user's shell as an ordinary shell, which won't do all that work (see the link).

su always starts a new user session, reauthenticating the user as though they'd logged in from scratch. Any environment variables from the outside are cleared out in the inner shell.

So all together this equates to the command (supposing your username on the machine is also "blunderboy"):

su - blunderboy

which means "reauthenticate a new session as blunderboy and run a login shell".

The overall result of this command is to start a new session as though from scratch, as the current user. Why bother to do that? Likely to pick up on some side effects of su: in particular, because it starts a new user session, it will pick up on any changes to the groups the user is in, as well as other user or permission changes, and and environment variables are all reset. Ordinarily, adding a user to a new group won't take effect until they log in again¹: su counts as logging in again for that purpose, so the shell it starts will have the new group active.

¹ This is a simplification, but true enough for this explanation.