As an illustration from my own practice, I wanted to give the audience in a lecture on cognitive biases a tangible experience of the ideas behind the famous “debunking” of the even more famous phenomenon of the Dunning-Kruger Effect (DKE).
So I called Claude to the rescue and asked it to put together a simple simulator that would intuitively demonstrate how the typical DKE pattern can arise from measurement noise / regression to the mean, boundary constraints, and universal optimism bias (also known as the Better-Than-Average Effect) - without invoking any “dual burden of incompetence.”
Within two or three iterations, in about 5 minutes total, the simulator was good enough to use. You can check it out for yourself here.
For anyone who teaches, lectures, or explains abstract ideas for a living - or just for fun - this kind of workflow is becoming hard to ignore.
If you’ve come across or built simple simulators, demos, or visualizations that made a hard concept finally click, feel free to share them in the comments 🤓
Related notes
- Testing my GenAI skepticism
- Does GenAI make me a better (more rational) thinker?
- Agentic AI for visual data exploration
- Loftus & Palmer 2.0: Replicating human bias in AI
- NetLogo: Don’t tell me, show me
📄 Read the original post with full outputs on my blog.