Blog
Thinking out loud.
The Luge blog: why we're building AI agents rather than a chatbot, how a team works with its AI colleagues, and what we learn along the way. Written by the team building it — not press releases.
-
The Luge method
One developer, a fleet of agents: backlog triage, tickets, tech watch and the nightly board sweep all run through Luge itself. Read off the live instance.
-
Delegate, don't operate
Driving an AI gesture by gesture is still doing the work yourself. Luge's founding principle: hand objectives to AI colleagues — with the four pillars that make delegation possible.
-
Why we built an AI employee, not a chatbot
A chatbot answers and forgets. An employee learns, remembers who taught it what, and carries the work through. The manifesto behind Luge, in five ideas.
To get future posts without checking back: the RSS feed.