Ideas

The thinking underneath the work.

Field notes on what good AI means once it ships, how we build it, and how we keep the impact numbers honest. Positions we hold, grounded in the work that proves them.

Good AI: our positionPillarOur position on what good AI means when it is actually shipped, not slogans: built with the people it serves, honest, and consent-first.AI-native, not AI bolted onField noteWhy AI built in from the first commit behaves differently from a chatbot bolted onto a product that was finished before AI arrived.How AI actually helps conservationField noteThe useful role for AI in conservation is not surveillance from orbit. It is helping the people on the ground count, coordinate and keep working with no signal.Measuring real impact with softwareField noteImpact numbers are only worth publishing if they are true. How we log outcomes at the event and cite every figure to the database it came from.

Positions are cheap. The work is the proof.

See the workRead the impact