Judgment Is the Job Now
For most of the history of work, the hard part was making things. Judgment lived at the edges, squeezed in around the making.
That ratio has flipped. A lot of my work is now a few minutes of asking and a long stretch of deciding what to do with what comes back. The making got cheap. The deciding didn’t, and the brain decides with shortcuts.
Your Brain Is Also a Model
Psychologists have been cataloguing those shortcuts for about fifty years, and the Cognitive Bias Codex maps 188 of them. Most of the time they’re the reason you get through the day. But each one misfires in a predictable way, and a bug you can predict is a bug you can catch.
AI leans on exactly the ones that matter:
- Automation bias. We trust answers from a machine and check less. A language model is the most persuasive automated system we’ve built, and it never hesitates.
- Fluency. Well written feels true, and models write just as smoothly about what they got wrong.
- Anchoring. The first draft quietly sets the range of what you’ll consider.
- Confirmation bias. Ask a leading question and you often get your own view back, better argued. A mirror that talks back.
It cuts the other way too. My last post was about the reflex that dismisses anything AI made. That’s a bias as well. Over-trusting and over-dismissing are the same mistake: letting a shortcut answer a question you should have looked at.
The Soft Skill That Doesn’t Expire
Prompting tricks change every few months. The map of human shortcuts hasn’t changed in fifty years. Anchoring worked on people in 1974, and it will work on you tomorrow, whichever model you use. That’s why knowing your biases, and having habits that catch them, may be the most important soft skill of this era. It’s the one part of the loop nobody can automate: deciding whether to believe the machine.
Knowing a bias’s name won’t make you immune. We spot biases in others far more easily than in ourselves. What helps is feeling one happen to you, and a habit that catches it. It doesn’t take much: in one study, a single short training session reduced several biases, and the effect was still there two months later.
So I Built a Map
The Cognitive Bias Explorer puts all 188 on one interactive wheel, each with an everyday example and how to spot and counter it. There are fifty-four thirty-second videos, ten experiments that let you feel a bias happen (spin a wheel, then watch a random number drag your estimate), a quiz, and eight habits that catch most of the map. Mark the biases you recognise in yourself, and it tells you which habits to start with. It’s fact-checked, in English and Greek, and yes, built with Claude. Look before you judge.
Don’t try to learn 188 biases. Pick one habit. A good first one: before you read an answer, decide what would change your mind.
The machines get better at the work every month. Deciding what to trust stays ours, and it runs on an old brain with known bugs. The least we can do is read the map.
Your brain takes shortcuts. Here is the map.