Active Inference
Brains generate predictions and act to make them true, minimizing surprise (free energy).
ELI5 — simple story
Your brain is always guessing what will happen next and moves your body to make those guesses come true as cheaply as possible.
Building on previous concepts
Taṇhā showed us the grab reflex. Active inference explains WHY we grab - we're trying to make predictions come true.
Concepts you've learned so far:
Felt experience
What does this concept feel like in the body and behavior? Read, notice, then try.
Active inference describes a continuous loop: the brain predicts what sensations should arise, compares these predictions with actual input, and acts to minimize any discrepancy. This creates a unified framework where perception and action work together to minimize surprise.
Free energy is a measure of surprise or prediction error. The brain constantly works to minimize this by either updating its predictions (perception) or changing the world to match predictions (action). This principle explains both learning and behavior.
Mechanism
What the animation teaches: This loop runs continuously: predictions drive actions, actions change sensations, sensations update predictions. Notice it's a closed loop - there's no 'start' or 'end', just continuous cycling to minimize surprise.
Active inference operates at multiple levels simultaneously. Higher levels make more abstract, longer-term predictions, while lower levels handle immediate sensorimotor control. This hierarchy allows for both rapid reflexes and complex planning.
Try it
Reflection: Before reaching for an object, pause and notice your predictions: where is it, how heavy will it be, how your hand should move?
This exercise builds awareness of the constant predictive processing underlying simple actions.
Thought experiment: Think of a recent surprise. How did your brain quickly update its predictions to accommodate the new information?
Understanding how surprise drives learning illuminates the active inference process.