Premise
Most data work in AI is organized around a bad assumption: that human labor should be cheap, replaceable, and hidden beneath the glamour of the model. Rubric exists in opposition to that.
Approach
The thesis is that annotation can be designed as a learning system. People enter as learners, work against structured knowledge, and become more capable as the system improves.
The data gets better because the people get better. That is the standard the branch should be held to.
Rubric is still vision-stage, but it is one of the clearest expressions of what the wider Wiyc stack is for.