Watching AI write code changed how I review it
Cold diffs are the worst way to review AI-written code. The agent's trajectory, which files it read, what it ran, how many retries, tells you where to actually look.
Termi Protocol Blog4 min read
The Termi Protocol
6 stories
Six AI agents, one codebase: what a normal day looks like now
Lanes in the morning, checkpoint diffs at night. A day in the life of a codebase worked by several CLI agents in parallel, and the rules that keep it from turning into chaos.
Termi Protocol Blog3 min readManaging AI coding agents is becoming a real job
Less typing, more directing. The brief, delegate, inspect, unblock loop looks like management, and the skills that transfer are not the ones people expect.
Termi Protocol Blog3 min readA week of barely supervised AI agents: what broke, what surprised me
I gave agents a side project, turned approvals mostly off, and queued tasks for a week. Scope creep, a quietly weakened test, impressive long runs. Honest notes.
Termi Protocol Blog3 min readVibe coding is not no-code
No-code hides code from you by design. Vibe coding produces more code than you can read. They demand opposite skills, and confusing them is how people get hurt.
Termi Protocol Blog3 min readWhy I built a 3D room for my AI coding agents
The work went invisible, so I gave it a body. How a weekend prototype of a robot walking to a desk turned into The Termi Protocol.
Termi Protocol Blog4 min read
Showing 6 of 6 stories
You're all caught up.
AiByts Weekly Brief
Get the top AI news and tools straight to your inbox every week.
Weekly · AI news
Join 15,000+ AI enthusiasts