Exercise 1: Prompt Engineering vs. Loop Engineering — Possible Solution
====================================================================

WHO READS EACH RESPONSE UNDER PROMPT ENGINEERING
------------------------------
Per this chapter, prompt engineering optimizes a single exchange: you
write an instruction, the model responds, YOU read the output, and if
it's wrong, you write a better instruction. The human is the one
reading every individual response and deciding what happens next -
"you hold the pen for the entire conversation."

WHO READS EACH RESPONSE UNDER LOOP ENGINEERING
------------------------------
Per this chapter, loop engineering optimizes a system that runs that
same conversation on your behalf, repeatedly, "without you reading
every single reply." The loop itself - not the human - processes each
individual response and decides the next action, based on a goal-
gather-decide-act-observe-repeat cycle.

WHY THIS IS THE ACTUAL DIFFERENCE
------------------------------
Per this chapter, the shift isn't about the underlying model or the
individual quality of a response - it's about where the effort of
reviewing responses goes. Prompt engineering keeps a human as the
constant reader/decider; loop engineering designs a system that takes
over that reading/deciding role, with the human's effort moving
"upstream" into designing the loop itself.

WHY THIS WORKS AS AN ANSWER
------------------------------
It correctly identifies the human as the one reading every response
under prompt engineering, correctly identifies the loop/system as the
one reading responses under loop engineering, and correctly frames
this as the actual distinguishing feature between the two approaches.