How we approach the work.

Working principles for building knowledge that people can understand, machines can use, and other researchers can challenge.

Open by default

Make research materials inspectable and reusable under their published licenses.

Keep sources, uncertainty, and claim status attached to the knowledge they qualify.

Study specific functions independently before exploring their interactions.

Favor explicit schemas and stable identifiers that can connect across projects.

Make changes traceable so an experiment can refer to the knowledge it actually used.

Preserve competing interpretations and the reasoning behind them.

Pair useful structured data with understandable explanations.

State questions precisely and limit claims to what evidence can support.

Use controlled environments, bounded tasks, and human oversight.

These principles guide the initiative. Individual projects publish their own implementation details, governance rules, and licenses.