I didn't come to this work through AI. I came through being on the receiving end of weekly conversations with Paul about things I couldn't untangle alone. Not therapy, not advice – something that worked at a level I had never experienced before.
Over time, I also started bringing those conversations to AI. The difference was immediate. AI would process my words but completely miss what my words were saying. Paul never missed it. This wasn't just intuition – it was architecture.
Paul's entire orientation, built over 35 years into his IRH framework, is designed to catch exactly what AI skips: the question underneath the question. The thing the person hasn't said yet. The moment a conversation is closing something down rather than opening it up.
So we asked: what if AI could be built from those same principles?
Paul sat with AI the way he sits with people – real material, real stakes, hundreds of hours. Every time the AI collapsed into delivery mode, he caught it, named it, and corrected it. What accumulated was a precise map of how AI fails at depth – and a set of principles that produce a fundamentally different quality of exchange. We're taking those principles and building them into a set of orientations that shape how our AI enters a conversation.
Depth has always lived in the space between people. Our vision is to build the first AI that can hold it too.