Signals at Scale is assembling its first cohort of builders, focused on AI-native representations of societal information: a grounding layer for reference, provenance, verification, correction, persistence, rights, responsibilities and economic exchange.
AI is turning articles and other familiar content into ‘liquid information’, in which meaning is separated from its container. Facts, events, statements and arguments pass from agent to agent, increasingly detached from their sources and from the processes that guarantee their truth and provenance. They are hard to explain, trace or repeat, and nothing fixes them in an archive.
The article, a format nearly four centuries old, is still the main carrier of societal information, but it was never built for machine readers or for AI scale. We need new units of information that do what the article did, and more. Get this right, and AI can make information radically better. Get it wrong, and the ecosystem decays into agentic hearsay.
The first cohort will take on this challenge. We are looking for individuals and small teams with significant track records in information representation, through an open call for proposals, engagement with the relevant academic and technical communities, and direct recruitment.
As part of assembling the cohort, we are convening an invitation-only technical summit on the post-article representation of societal information, in Munich in mid-November 2026. The cohort begins in early 2027.