Coding with LLMs makes me think more

It's not uncommon for devs to just start coding and spending little to no time on designing the solution and thinking about the big picture. When working with AI-assisted tools like Cursor, I'm forced to think more about what I want the agent to do, so that it doesn't get lost.

The tangible output of better thinking is better context for the LLM (see Context compression for better code generation)