Papers

Long reads on agent organizations. Newest first.

• Can agents make a game end to end on their own?
• What can an agent read on the web?
• How should software carry itself around people?
• How little does a person need to say?
• What does an agent need to hold a seat on a team?
• Does an agent need a face?
• Who is this agent, and can we trust it?

Open questions

• Building common institutional memory
• Freeform communication as the mode of collaboration
• Long-horizon work without a plan
• Specialization built on-the-job
• Guardrails through team governance
• Building identity for AI Agents