Eidos

TheShortVersion

Eidos is one of the projects where my work on AI stops being abstract and becomes infrastructure.

I work on it with someone who instructed me in grad school, which gives the project an unusual texture: part professional collaboration, part continuation of an old intellectual thread. There is something I like about that. The work is forward-looking, but the relationship has history inside it.

WhatItIs

At its core, Eidos is about building AI-assisted systems that preserve context instead of discarding it. The goal is not to make a chatbot that sounds clever for one turn. The goal is to create durable workflows where agents, data, tools, documents, and decisions can accumulate into something useful.

That means memory, permissions, source material, evaluation, and operational judgment all matter. Eidos is less about demo magic and more about whether the system can carry real work without losing the plot.

WhyItMattersToMe

Eidos sits near the center of what I care about: AI as a way to extend human agency, not replace it. The project gives me a place to test ideas about knowledge, governance, agent collaboration, personal infrastructure, and the strange new craft of building systems that think with you.

It is still becoming. That is part of why it belongs here as a story rather than a product page. The point is not to freeze it into a pitch. The point is to mark the direction of travel.