One thing that’s jumped out to me as I use more and more AI tools, especially for coding: using pen and paper is even more important now that it was before.

It’s always been important to work out things on paper. The Feynman anecdote on written notes comes to mind:

Weiner: Well, the work was done in your head but the record of it is still here.
Feynman: No, it’s not a record, not really, it’s working. You have to work on paper and this is the paper. OK? (source)

Initially, using an LLM as a sounding board to work things out is tempting, and feels like an easy way to start. But I find that it’s essential to work the basics of what I want out for myself first before going to the LLM. Without some clarity in my mind, it’s easy to be led into a local maxima, even when it’s just me and the LLM in the room.

So now I find myself drawn to pen and paper even more than before. Work it out for yourself, then go to the AI.