A Trick of the Mind summary

To put it simply, perception just isn’t enough. The brain also needs a manual for doing – taking action. So, again, much like a scientist, it builds models of cause and effect, which is essentially a model of you. Each action you take – flip a switch, crack a joke, move a finger – is a tiny experiment, and the results feed back into a theory of what you can influence.

And yet, just as in the laboratory, our hypotheses can backfire. Sometimes we’ll exude confidence in a situation that goes completely out of our control. Other times we’ll feel completely helpless in circumstances that are completely in our control. How often have you pressed an elevator button with increased frequency or pressure even though you know full well it’s not going to have any effect?

Let’s move away from our internal inspections and look outward to others. Every day we’re attempting to decode other people’s intentions and emotions, even when their signals are ambiguous. A colleague’s silence, a friend’s sudden exit – are they annoyed? Distracted? What’s going on?

To cope, the brain runs what’s called a “Galileo maneuver”: it first points its instruments inward, using firsthand knowledge of how feelings drive our own movements, then projects those models outward to interpret others. Motion can tell us a lot: buoyant steps often signal joy, heavy movements sadness, sharp bursts signal anger. 

If we think of our brains as scientists running constant experiments, it makes sense that confidence grows with successful predictions. Research shows that early experiences of success and failure shape our internal models in lasting ways. Wins give us a boost and keep us engaged, while losses can cause us to shrink away. Yet persistence pays off: people who keep trying after setbacks often catch up with the so-called early winners. 

Building a self-model that emphasizes perseverance can require some effort. As we’ve touched on, people routinely misjudge their own behavior. Our self-monitoring system, known to scientists as metacognition, or thinking about thinking, has to deal with a lot of messy signals. It can be a challenge to know what to trust and what to ignore.

Dopamine has a certain reputation these days. Most of us have heard about this neurotransmitter and think of it as the brain’s pleasure chemical. But that’s only a small part of the story. The joy we feel when we get a hit of dopamine isn’t just related to the pleasures of hedonistic pursuits. It also comes down to our brains being wired for curiosity, and the joy of discovering and learning. One of our defining characteristics is our innate drive to chase understanding even when it has no obvious payoff. Why else build particle accelerators or spend evenings lost in philosophy?

What gets the pleasure center of the brain really going is surprise. Sure, we all like getting a reward, but the real turn on is found in the gap between expectation and reality – when there’s a prediction error and the reward is unexpected. If we can accurately predict when the prize will arrive time and time again, it becomes boring – we lose interest.

Earlier on, we mentioned how the models in our brain were similar to the models used by machine learning. The better the data sets, the better the output. And as artificial intelligence continues to improve, it may lead you to wonder, if machines can stitch together fluent sentences by tracking patterns, are our brains doing something similar? Are we also just prediction machines at heart?

It’s true that we put a lot of emphasis on language. When Google engineer Blake Lemoine had his first breakthrough with LaMDA, a large language model, he believed it had become sentient, simply because it was talking to him in such a clear and convincing way. But even though others were quick to argue that these systems are just elaborate pattern-matchers, their output can feel eerily human.

Throughout the Blink we’ve followed the “scientist in your skull,” watching how it steadies perception with theories. But models wobble. When the world grows volatile, the same machinery that guides us can lure us toward odd beliefs – including conspiracies. The issue here isn’t strange minds, but uncertain times. And when uncertain times lead to paradigm shifts, our models can really get unbalanced.

When our everyday world shifts from “normal” to being full of anomalies, our confidence cracks. Our expectations and predictions feel off. A paradigm shift forces the question of trust that our brain faces daily. Do we trust old experiences or new data? When should we remain stubborn in our beliefs, treating anomalies as chaotic noise and coincidence, and when should we be flexible, and update our beliefs accordingly?