Coming from a non coding background, I’m curious if anyone has some tips to limit the AI slop while using Claude code. Also open to other use cases with AI coding tools.
What works for me is being very specific about the requirements and tasks. I usually create a plan first, then break it down into smaller steps, and then pass it off with proper rules and context for each step. The more open-ended the request is, the more slop will be generated. Smaller, directly focused tasks, and explicit constraints seem to help a lot.