First, those
two loops are NOT two "versions" of the same knot ! They are
two different knots - as different as they could possibly be !

Starting from the second, its resemblance to the
Ampersand TIB bowline is obvious - although, to
my aging mind, at least, it has this drawback of the "strange" path the second leg of the collar follows, as it does NOT go through the nipping loop like the first one, as I use to expect... My mind has ceased to evolve !

I still find it
easier to tie a bowline-like eyeknot where both legs of the collar pass through the nipping loop ( from the same, or even from opposite sides, as in the Mytrle-collar loops ).
However, the great thing with the
Ampersand TIB eyeknot, which is absent here, is that it is a TIB eyeknot !

Having a loop that is PET
AND TIB is a GREAT convenience,
which nobody can deny !
Now, the first loop is an altogether different animal. It is also not a TIB eyeknot, and I do not see how one can change it to become one either. It
looks complicated, and it
is !

However, this ( counter-clockwise, 180 degrees ) "twist" of the eye during its tying, right after the second stage, is an
ingenious move ! ! ! I had not seen it before, and, of course, I had not thought of trying something like this, never ! It is a very simple yet potentially very productive move, which can generate many new eyeknots !
Twist the eye of the to-be-tied loop, in one stage of its tying, is a new idea, which deserves further exploration. Who knows how many more loops you can tie by implementing this...Moreover, IFF those loops just happen to be TIB, too

, it would be a great thing, indeed !