I used to walk around thinking about code. I could "see" the code in my head. It was pretty cool. If there was a bug I could often click straight into a specific file and fix it with no other context, like a type of muscle memory. Now I think in prompts. I'll be cooking dinner…
I used to walk around thinking about code. I could "see" the code in my head. It was pretty cool. If there was a bug I could often click straight into a specific file and fix it with no other context, like a type of muscle memory. Now I think in prompts. I'll be cooking dinner and recall something codex said to me, then imagine myself typing a prompt back. thinking in prompts seems strictly more embarrassing than thinking in code. not sure if I'll ever go back though
discussion1 selected reply
Morris describes moving from a mental map of files and code to thinking in prompts. That abstraction can widen the work an agent handles, while making codebase-local memory less central to the developer’s own reasoning.

I feel it’s harder to build a mental map of the software in my head because I mostly read diffs and AI responses/reports, rarely read full code. I have to explicitly ask AI to generate design diagrams to build a mental map. Thinking in prompts is real too.