Stop Tweaking Prompts: Why Treating Claude Code Like a Coworker Finally Made AI Click
Fit_Needleworker_542 · reddit · 2026-09-26
The author describes their breakthrough after months of frustration with GPT, Claude and Claude Code: they were obsessing over prompts, tweaking and re-asking until managing the AI took longer than doing the work.
What changed was treating the model less like Google and more like a coworker:
- Give Claude Code real project context, let it explore the codebase and understand the goal
- Let it make changes, run tests, and fix issues in a back-and-forth loop
- Shift the question from "what prompt should I use?" to "what can I make this do for me?"
They also cite someone who used Claude Code to build a 9,700-page property website, and credit content by Trent Calloway with helping reframe how they use AI to automate whole tasks instead of manual grind.
More from coding & agent
- Andrew Ng: AI engineering tactics must calibrate to project stage, from evals to architecture — DeepLearningAI · 2026-09-26
- Running multiple Claude/Codex sessions on NPM? Switch to pnpm and save tens of GB — altryne · 2026-09-26
- One Skill for World-Class Agent Research: Multi-Source Verification and Skeptic Voting — EXM7777 · 2026-09-26
- AgentStuff Launches: Plan Once, Let Claude Code and Codex Ship Projects in Parallel — tedddyoweh · 2026-09-26
- Simple KJnodes workflow adds bounding-box control to Krea 2 text-to-image — Mulksky · 2026-09-26
- Agent curates monthly clothing rental box, turning hours of scrolling into minutes — nikhilaravi · 2026-09-26