Context, Communication & Community
A small community became a surprisingly useful environment for exploring AI-mediated communication.
The usual problems were already there: different languages, cultures, time zones and communication styles. I happened to have the technical background — and a group of people open enough to experiment together.
So we started exploring what happens when everyone can use AI as part of their communication.
I had the opportunity to build an environment around a private server and Matrix, connect my own AI agents to it and make some of those capabilities available to the community. People could use AI themselves, interact with shared agents, or simply benefit from support happening in the background.
That made the interesting part visible very quickly.
Translation was not the hard problem. Context was.
When different people use AI to formulate, translate and interpret communication, the interaction is no longer simply human-to-human. AI becomes part of the communication layer. Intent, individual communication styles, shared history and the quality of context suddenly matter enormously.
It was also a practical lesson in trust: a convincing AI interpretation is not necessarily a correct one. People need to understand where AI is involved, what information it has, and how much confidence an output deserves.
For me, this is what exploring AI looks like.
Find an appropriate environment. Build enough to make the idea real. Let people use it. Observe what actually happens. Then take those lessons back into architecture, user experience and the future of work.
And sometimes one experiment creates the next one.
A later idea to make some music for the community eventually became HooX.