Technology

An order-of-magnitude estimate of what it costs to build and serve the claim graph, doing it right. The cost is dominated by LLM inference; a frontier-quality build lands in the high single-digit millions, concentrated in the broad middle tiers and the ingestion firehose rather than the flagship claims.

Compute is the scarce resource behind the claim graph. How to decide which claims get mapped, assessed, and re-assessed — a distributed, market-like approach with cost/value estimation and grantmaker agents.

Everyone agrees AI agents are a big deal; no one agrees what "agent" means. A taxonomy of the word's senses—taskbots, chatbots, drop-in remote workers, assistants, and freebooters—organized along two axes: who the agent serves, and how much agency it has.