Human in the Loop

Most people meet AI already finished. A model arrives, it behaves the way it behaves, and the only decision left is whether to use it. Epic Minds exists to put a community back into that decision.

What we do

We help a community stand up its own small AI cluster — modest hardware, owned locally — and we teach the people in it to run the thing. Not as a demonstration, but as something a school, a co-operative or a neighbourhood keeps running and keeps shaping.

On their own, those clusters are useful. Connected, they are something else: many small groups, each deciding what their AI ought to be good at, pooling that direction rather than waiting for a vendor to choose on their behalf. That is the distributed part, and it is the whole point.

Why it matters

A model reflects whoever shaped it. When that is only ever a handful of companies, everyone else lives with decisions they had no part in making. Running it yourself changes who gets to answer the question of what it is for.

Human in the loop usually means a person checking a model's output. We mean it a step further out — a say in what the model becomes, not just a veto on what it has just said.

The school technology work continues

Epic Minds started in classrooms, helping schools make their technology an asset rather than an obstacle, and that work has not stopped: Google Workspace, Microsoft 365, the learning management system in between, and the Llamas Toolkit.

It is also where the rest of this came from. Schools are where it became obvious that the people using a system are almost never the people who decide what it does.

Talk to us about either.