Ah ha! there. Hard to explain why this is cool to me. Back in
   1990/91, I studied a bit about neural networking (the stuff of
   Artificial Intelligence), which was based on a very old model of
   "how neurons might work" from back in the 1940s/50s based on
   weighing perceived significance on edges connected to nodes.
   Wait, no, easier: "Why is this connection important? I'll give
   it a score of 58%" "Oh wait but no, this thing over here is ALSO
   important with a significance of 80%", "CRAP! I'd better update
   that 58% and make it a 27%"
   One of those "everything affects everything else" type things.
   Anyway, it works *REALLY WELL* but it's _slow_ as heck if that's
   ALL you're using to sort out the world. Thousands of tries to
   figure out "simple things" when used on computers. [still
   awesome of course] Anyway, 2nd line in this, which is about the
   Event Horizon Model - same theory used to explain why we forget
   things passing through doorways... meant something to me. While
   some form of "weighted importance" may be true for sorting out
   long term memory when it enters working memory... [gotta throw
   out the expectation of finding pickles or poop when your mind
   pulls forward smells of pickles or poop when someone gives you
   bad news because it's irrelevant] - it DOESN'T account for
   stepping through life from event to event, why we forget
   someone's name we *just met* as we go into the car to leave, or
   a ringing phone can make us lose our "train of thought" so
   quickly. Weights are nice for pulling forward old memories for
   filling in gaps of knowledge about the present moment - but not
   a very efficient system for juggling working memory. Too big and
   clumsy. Working memory is a very speedy sorter, and is agile and
   goes from event to event, resetting itself whenever SOMETHING
   big enough changing, tossing the old working memory in
   compressed form back into Long Term Memory.... or maybe some
   kind of inbetween memory. [these are conceptual "buckets" - not
   tangible ones, I don't think] Oh, and #5 is interesting: It's
   the reason why we can't find the car in the parking lot: It's
   because EVERY MEMORY of EVERY car in EVERY parking lot is
   overlapping with each other.... and it's really hard, especially
   when it's the SAME parking lot, different place, to distinguish
   one event from the other... because they're all practically the
   same every time.[1]event-horizon-model

References

   Visible links
   1. http://icopiedyou.com/wp-content/uploads/2015/08/event-horizon-model.jpg