I’ve spent 3 months teaching ChatGPT to think like a behavioural scientist.

Here’s what I believe:

By the end of 2025, it’ll be better at behavioural science than me.

But only if I prompt it right.

So I gave it the same training I gave my own mind.

Learn when to diverge.
Learn when to converge.

That’s the Double Diamond process.

It changed how I write.
It changed how I design strategies for behaviour change.

Now it’s levelling up how I train ChatGPT.