Glossary · Term

DeLM

← all terms

Definition

Plain language

A way to coordinate many AI agents through a shared, fact-checked bulletin board instead of routing everything through a manager agent.

As stated in the literature

Decentralized Language Models — a multi-agent architecture pairing a pool of parallel agents, a verified shared context with hierarchical summaries, and a task queue, replacing a central controller with admission-gated shared state.

Also called: Decentralized Language Models

Why it matters: It lets many agents collaborate without a central bottleneck while keeping the shared workspace trustworthy through fact-checking before anything is written.

For example, ten agents working on a report each post their fact-checked findings to one shared board, and anyone can read the running summary instead of waiting on a single manager to relay updates.

Heard on the show

“What does the system — they call it DeLM, for Decentralized Language Models — actually look like?”
Episode 130 — Why AI Agents Coordinate Better Through a Shared Board Than a Boss

Mentioned in 1 episode

  1. 130
    Why AI Agents Coordinate Better Through a Shared Board Than a Boss

Related terms