MoMo: Building an AI Companion in Public

MoMo: Building an AI Companion in Public

The first session ended with a challenge worth taking seriously. If AI offers artificial connection, emotional chewing gum that lets people skip the friction real relationships require, then anyone building in this space has to answer for it. This session I'm putting my own work in the room and letting you do exactly that.

I'll walk through MoMo: what it is, what we have actually built so far and the architecture behind the part that matters most, which is memory that holds across months rather than resetting every session. Then the harder half. The technical problems we have not solved and the psychological ones, including where the line sits between a tool that helps someone notice a pattern and a tool that quietly becomes a substitute for the people in their life.

I am not presenting a finished thing and hoping for a fruitful discussion and engagement.

Roughly half the time is presentation, half is open. Whatever the conversation turns into, we follow it.

New format starting this session: one topic, one presenter, going deep instead of wide. Oleg is up next with his work on agent identity and the presence test.

No preparation needed. Bring your curiosity.