Ah yes, I do like that we no longer automatically import/1 at the same time!  That was being pesky on several fronts in one of my pack's code.  But wasn't sure how to clearly I could argue for the new behavior.

I'll still try to describe as it shows my support:

if a  autoload defines  other_guy:name/5 ..  (and lets say DWIM fails)   If I ask for

  myself: (....
    clause(name(.....),true),
    .... )

  I could have some errant other_guy:name/5 taking over my module before I got to my own   :- myself:dynamic(name/5).    In the file.  (My mistake surely, and I think even a warning happens) but still it "sometimes it I get away with it", because  my  :-dynamic(name/5) came early enough and I did not know..  But then am confused since the rest of the modules in my pack are using the an unexpected name/5.  ( the other side of the coin is if autoloads are supposed to be inherited from  'user' or 'system'  (thus they might still import to those places)  but they would still retain their definition modules