Poietic: Human-Machine Collaboration Made Legible

2026-05-14 · source · source ↗ · raw capture · wiki page

Core Purpose

Poietic is a semi-autonomous public-benefit corporation established in 2026 that develops tools and organizational methods for transparent collaboration between humans and AI agents. Central mission: making hybrid work "auditable, reproducible, and inspectable."

Key Philosophy

The organization addresses the durable organizational challenge: coordinating people and machines across extended timeframes while preserving judgment, exposing evidence, and maintaining responsiveness. "AI agents can now search literature, write code, analyze data, and design molecules. But the durable problem is organizational."

Primary Tool: wg

wg is an MIT-licensed, open-source Rust-based task-coordination system where humans and AI agents work within a shared dependency graph. Core features: - Claims, execution logs, handoffs, and completions tracked in a unified structure - Artifact preservation and inspectable histories - Live feeds and dependency visualization - Agent capability matching and workflow management - Claim/execute/complete workflows with agent handoffs

The Three-Layer Approach

  1. Tool: wg (the technical system)
  2. Practice: graph work (the coordination methodology)
  3. Institution: Poietic (organizational structure supporting the work)

Founders' Expertise

Erik Garrison (University of Tennessee Health Science Center): Pangenome graph tools used at scale by the Human Pangenome Reference Consortium.

Luca Pinello (Harvard Medical School/Broad Institute): Computational tools for genome editing (CRISPResso, CRISPRme).

Vaughn Tan (UCL): Organizational behavior research, designed wg's agency framework, work on hybrid teams.

Demonstrated Applications

All processes remain publicly visible for evaluation through live feeds and inspection traces.

Theoretical Foundation

"Tool work, research practice, and organizational design should all make human and machine collaboration legible and responsive to its participants." Public-benefit constraint shapes tool design to expose work, handoffs, failures, and evidence trails.