hey all, i’ve been experimenting a lot with coding agents and a different way of working with them
about 6 months ago i more or less stopped reading the code they write, not because i trust them blindly, but because with the right process i usually don’t need to
i’ve shared my approach on github and have shown it to a bunch of people across different roles. happy to help if anyone wants to get more useful output from agents without having to micromanage every step
if you want to chat about it, just ping me
Connect with me on LinkedIn - linkedin.com/in/marcocello
Harness engineering, building things, beer, muay thai
