id: 2151,4C2
title: Doom loops are a mutually negative feedback loops between learning loops.
date: 2022-01-01

Doom loops are a mutually negative feedback loops between learning loops.

They occur when one learning loop sees a negative result that its beliefs anticipated which further advises doubling down as the feedback, while a separate learning loop sees a similarly negative outcome with the same conclusion. A good example is taking things away from a dog that has cause to believe that it needs to defence what it has. Opposite of a delight loop’s positive reinforcement.[^G1]