Accuracy

What does happen when the camera is broken? How do we know to fix it?

Because models are simplifications of the world, they are wrong all the time: in big and little ways.

When the model is wrong, you experience a familiar feeling called “surprise’ (or, in scientific terms, “prediction error”!

Think of what happens when someone you know cuts their hair — the first few times you see them, your model of how that person looks is wrong, and it takes time to adjust. “I just can’t get used to you with short hair!”

Importantly our brains don’t just live stream information from your senses – that would be way too much information to process! Instead they use a similar mechanism to what streaming services use. They only transmit what’s changed from the prior model of the world. Only surprises or prediction errors get transmitted to our experience – everything else we experience is actually our brains’ model of the world.

Those errors tell us – “wait, this model is different from what we’re seeing out here. We’ve gotta check this out.”

Solution: Prediction errors allow us to constantly improve our models over time. The brain is trying to build you a perfect model of the world so you can perfectly meet your needs at all times. Neuroscientists believe that “minimizing prediction error” or “minimizing surprise” is literally the prime directive of the brain.