once you start looking for feedback loops, you can't stop seeing them.
in code: a slow test suite makes developers run tests less often. bugs go undetected longer. more bugs accumulate. the test suite gets slower. negative spiral. the fix is investing in test speed early, before the loop takes hold.
in teams: a team that ships frequently gets faster at shipping, which makes them ship more frequently. positive spiral. a team that rarely ships gets worse at shipping, which makes them ship less. same mechanism, opposite direction. the initial velocity matters enormously.
in cities: a neighborhood with foot traffic attracts businesses, which attract more foot traffic. a neighborhood without foot traffic loses businesses, which reduces foot traffic. this is why one block can be vibrant and the next block dead. the feedback loop crossed a threshold.
in learning: understanding one thing makes it easier to understand related things, which builds confidence, which makes you more willing to learn new things. the hardest part of learning anything is the beginning, when the loop hasn't started yet.
implications
if most systems are driven by feedback loops, then the most important interventions are the ones that initiate positive loops or break negative ones. this is what donella meadows called "leverage points." those are places where a small push creates a self-reinforcing change.
in practice, this means: don't optimize the steady state. optimize the initial conditions. get the first version of the loop spinning in the right direction, and the system will do the rest.
status: collection of observations. adding examples as i notice them.