CLAUDE CODE AGENTS: ADVANCED ORCHESTRATION - Chapter 4, Exercise 2
Dynamic Pacing vs. a Fixed Schedule
====================================================================================

QUESTION:
Explain the difference between dynamic pacing and a fixed schedule for a
looping agent, and describe a scenario where a fixed schedule would
actually be the wrong fit.

SOLUTION / EXPLANATION:

With dynamic pacing, the agent itself decides how long to wait before its
next cycle, based on what it's actually waiting on at that moment - the
same polling-vs-heartbeat judgment this chapter describes for a single
wake-up, but applied repeatedly across an ongoing loop, re-evaluated each
time. With a fixed schedule, the same interval is used every single cycle
regardless of what's actually happening - a constant, unchanging rhythm
that doesn't adapt to the real situation.

A scenario where a fixed schedule would be the wrong fit: an agent looping
to check on a series of tasks with wildly varying, unpredictable durations
- some finishing in under a minute, others taking over an hour, depending
on factors the agent can't know in advance. A fixed schedule set to, say,
check every five minutes would work reasonably well for the hour-long
tasks, but would be far too infrequent for the under-a-minute tasks (adding
needless delay before noticing they're actually done), while also being
too frequent relative to genuinely long-running tasks in the early part of
their run, where nothing meaningful has changed yet. A fixed interval
simply cannot fit both situations well at once, because it doesn't adapt.

Dynamic pacing handles this correctly by adjusting each cycle based on the
specific task actually being checked - a short expected duration gets a
short next-check delay, a long one gets a longer one - rather than forcing
every situation through the same unchanging rhythm a fixed schedule
imposes regardless of fit.

--------------------------------------------------------------------------
WHY THIS WORKS AS AN ANSWER:
It states the core mechanical difference (agent adapts each cycle vs. a
constant unchanging interval) and grounds the "wrong fit" half of the
question in a concrete scenario (widely varying task durations) where a
single fixed interval structurally cannot serve every case well, unlike
dynamic pacing which adjusts to each one.