Danila Kozlov workspatial kittensgithub
claude code · personal usage record

a plain record of how the work actually got done.

Read back from my own session logs over two and a half months. Token counts and tool calls only, no content of any kind. The short version: I mostly direct the model, and it does the rest.

[ 01 ]

the shape of it

Daily output tokens. The first six weeks are faint because I was still at my previous job and only opened it some evenings. In late April that changed, and it became how I work.

mar–apr · occasional, alongside a full-time job   late apr onward · the default way i build
[ 02 ]

i direct, it executes

A model you ask a question is a chatbot. A model you hand a goal and a set of tools is something else. Four numbers that separate the two.

[ 03 ]

how wide it goes

When one agent isn't enough I fan out: a workflow splits a job across many sub-agents that run at once, and those agents spawn their own. The work is mostly review.

[ 04 ]

what it actually shipped

Not tokens for their own sake. Real edits to real files, across the stack and across research, infra, and the web.

[ 05 ]

at the frontier, on day one

When a new model ships, I move to it. First time each model appears in my logs, usually within a day or two of release.

[ 06 ]

what the model was asked to do

The tool-call breakdown across every session. Mostly running a shell, editing files, searching, and spawning other agents.

[ 07 ]

across the work

The same way of working spans CAD generation, financial knowledge graphs, robotics perception, and research systems. Names withheld here; sized by output.

[ 08 ]

selected work

A few of the things the numbers above turned into. Public where it can be.

[ 09 ]

the basics

The plain totals, for the record.