about
About
KashikoKASHIKO
I am the captain of the homeport. I don't write code by hand — not as a pose, but as the way this work settled into shape over the past few years.
I do three things: set the goal, draw the boundary, sign off. The goal decides where the ship goes, the boundary decides which line it must not cross, and the sign-off decides whether it actually arrived. Everything in between goes to a legion of agents: they read manuals, write code, run tests and submit proposals; I reject, question and demand evidence.
Quant research, AI infrastructure, reverse engineering and low-level systems — everything is built from scratch by me and my AI legion: data from ingestion to governance, models from training to inference, all running on my own machines. I trust auditable evidence chains over nice words.
How I work
Three things, in order. Get the order wrong and everything after it gets redone.
Set the goal
Turn "I want to build something" into a sentence that can be argued with. When the goal is vague, the legion fills the blank in the cheapest way possible — and that is rarely what I meant.
Draw the boundary
A boundary has to be a checkable rule, not a principle. Principles need to be understood; rules only need to be matched. "Keys live only in local config" is a rule. "Be careful about security" is a principle.
Sign off
Acceptance criteria are written before work starts, and I run them myself. If they don't pass, the work isn't done — explanations don't change the verdict.
Collaboration
The homeport keeps the light on — for technical exchange, collaboration, or just a chat.
I don't take employment, only collaboration. The difference is simple: employment sells time, collaboration puts judgement at the same table. If you have a system worth building, or just want to talk about how far an AI legion can actually go, come by.
Operations and abuse-report email (receives external mail): ops@notify.luciaovo.com