Three videos I’ve been watching recently that are worth your time.
Claude Code + CMUX: The Ultimate AI Coding Terminal
Pairs Claude Code with tmux to build a fast, keyboard-driven AI coding environment entirely in the terminal.
Key takeaways:
- Using CMUX alongside Claude Code reduces context switching significantly
- The setup is fully keyboard-driven - no GUI needed
- Shows a practical workflow for developers already living in the terminal
- Good for anyone who wants to lean into AI-assisted coding without leaving their shell
OpenClaw Explained: Why the Hype is (Mostly) Wrong
A grounded, skeptical breakdown of OpenClaw and the noise surrounding it.
Key takeaways:
- Separates what’s genuinely novel from what’s marketing
- Argues the core claims are overstated relative to what the tech actually delivers
- Useful framing for evaluating any hyped AI release
- Worth watching before forming a take based on headlines alone
Stop Fixing Your Claude Skills. Autoresearch Does It For You
Makes the case that manually iterating on Claude prompts and skills is the wrong approach - Autoresearch can handle that loop for you.
Key takeaways:
- Autoresearch automates the trial-and-error of prompt refinement
- Frees you from the tedious work of manually debugging Claude skill outputs
- Shifts the mental model from “fix the prompt” to “define the goal”
- Practical demo showing before/after results with the approach