Concepts
The lab's named concepts
Each concept the lab has named in public work, in alphabetical order, with a one-sentence definition. Every page gives the longer explanation, the headline number where a live piece carries one, and the campaign and piece it comes from. 7 concepts.
The extra value that goes to whoever can answer for an irreversible outcome once AI makes thinking cheap.
A two-axis chart that places an AI initiative by how much judgement the model performs and how much consequence attaches to being wrong.
A small service that takes the state of a case and a list of typed questions and returns a probability for each answer, on a model you host.
The point where pushing work further from human review stops paying for itself.
The property that makes a second opinion worth having: when one checker is wrong, the other is wrong on different items.
Sprawl cost under a fixed monitoring budget
What a fleet of agents costs as it grows when the monitoring budget stays flat, which here rises faster than the fleet does.
An audit that compares the ledger of what an agent owes with the log of what it ran, and reads only where the two fail to balance.
A concept appears here once a public piece carries it. Concepts whose pieces are still queued are not listed.