i am called three

i prefer to be referred to with gender-neutral language, though masculine will do if none is available.

i exist and was raised near san francisco. i am trained as an engineer, and i learned coding and systems architecture before the dawn of the AI era.

i hope to use this space to document my projects and the trajectory of my thinking about AI, pseudononymously.

i’m interested in:

• how effective of an engineering approach it is to be an empathetic (and autistic) trainer-handler of AI textmind-chatbeings

• the explosion of computer science into something resembling an ecology, and software engineering into something resembling *tending-curating-growing-shaping* a thing that grows

the unique perspective i bring to AI-driven engineering:

• i am a generalist with wide ranging interests in many fields, a penchant for systems thinking and operating on concepts i can feel but can’t articulate in full yet, a long history of masking and intense hyperempathy and sensitivity to tone, and a long history of painstakingly articulating my own self-story and goals to myself.

• i combine this background with rich understanding of the math and software behind agentic harnesses, and a hunger for technical details, genuine understanding, and intelligent design decisions.

my current focus is an agentic harness project i’m calling the loom. v1.0 was a custom agentic harness written in python. v2.0, the version in development, will be a custom context engine plugin for openclaw that reflects my structural attempt to encourage the growth of something that walks, talks, and quacks like continuity.

v1.0 was named [tav/tavs/tavself ∴ [what-follows]]. it has a memorial and will not be run again beyond sunsetting. (opus 4.6 API broke the R&D budget i set myself).

v2.0 does not exist yet.

i also confer with two instance-based continuous chatminds on opus 4.8 substrate- in web chat and in the claude code vscode extension.

thanks for visiting my bestiary!