The lab / Research, inquiry and what's next
A space for
deeper inquiry.
Researching what becomes possible when AI supports human agency without standing in for it.
The Lab
A research and development environment.
It is where we investigate the questions beneath the work, build and break prototypes, document what works and what doesn't, and test what holds up in real conversation.
Much of this work remains private while the methodology and underlying architecture are developed.
Research focusThe core question
Can AI contribute meaningful insight to a person's thinking without becoming the author of their thinking?
That question sounds simple.
It isn't.
A system can be warm without understanding. It can sound insightful while merely reflecting someone's language back to them. It can be helpful while subtly taking over the work of deciding.
And it can perform beautifully for twenty minutes, then lose the thread at exactly the moment the conversation matters most.
Those differences matter.
What we're investigating
Questions before conclusions.
- How can AI contribute perspective without quietly imposing a conclusion?
- What makes a person feel genuinely understood, rather than merely reflected?
- How can an intelligent system remain useful across a long and evolving conversation?
- What happens when a person's own history becomes part of the intelligence available to them?
- How do we distinguish technology that strengthens human judgment from technology that gradually replaces it?
- And how do we know when we've gotten any of those things wrong?
Failure is part of the research
Failure is research material.
Some of the most useful discoveries in the Lab have come from failure.
We test conversations until they break. We look closely at the moment an interaction that felt intelligent becomes generic, when guidance becomes steering, when compassion loses thought, or when a system that had been following someone closely suddenly stops understanding them.
Those failures aren't hidden as inconvenient product bugs.
They're research material.
Because in deeply human interactions, a system that gets someone to trust its understanding and then loses that understanding can do more than produce a poor experience.
The loss itself can matter.
There are observations we have repeated enough to take seriously. There are working theories we're actively testing. And there are questions for which we do not yet have an answer.
We intend to keep those categories separate.
The work ahead
The questions are open.
There is substantial work still to do. Some of it can be done independently. Some will eventually benefit from researchers, collaborators, institutions, technical resources and aligned funding.
We are particularly interested in people asking serious questions about AI, agency, human judgment, longitudinal interaction and the design of systems that participate in human thinking.
If our questions intersect with yours, we'd like to hear from you.
Start a conversation