#2026

Agent Kombat

A small shell idea for making Claude Code and Codex argue over the same plan before I trust it.

#2025

Sorting Prompts - LLMs are not wrong you just caught them mid thought

LLMs are not wrong that often, you're just forcing them to answer mid-thought. Mike Caulfield has a clever techinque or prompt that will help you get much better results with your AI searching.

Build your own /init command like Claude Code

You don’t need Claude Code's `/init` to make repos AI‑ready; save this prompt and use it anywhere.

Three important things to get right for successful AI Coding

This is a post to help developers struggling with inconsistent AI coding results. There's just three things you need to focus on today: Context, prompts, and chunking. This post helps you understand how to effectively nail these three factors.

tags

kau.sh | Prompt-Engineering

Writing Letters About

Subscribe

Follow me via email, RSS, twitter or other places.