{"type":"post","id":"how-to-train-a-swarm-19-minute-read-4fb1ec","slug":"how-to-train-a-swarm-19-minute-read-4fb1ec","title":"How to Train a Swarm","description":"General Reasoning describes how reinforcement learning can encourage an agent collective to divide work, communicate when needed, and finish sooner.","retrieval_nugget":"General Reasoning describes how reinforcement learning can encourage an agent collective to divide work, communicate when needed, and finish sooner. Its RLCR recipe uses shared rewards and time pressure; the practical point is that parallelism must be rewarded through the training objective, not assumed to emerge from adding agents.","published_at":"2026-11-03","updated_at":"2026-11-06","record_date":"2026-11-03","date_kind":"published_at","topics":["agent-economics"],"entities":["General Reasoning","RLCR","DeepSeek-V4.1-Flash","Open-Reasoner-Zero"],"related_patterns":["agent-economics-moves-to-completed-work"],"novelty_label":"new_fact","what_changed":"Adds a training method for swarms: RLCR uses shared-policy collective rewards and latency pressure to encourage parallel work and coordination.","prior_refs":["https://newruntime.com/posts/yc-labs-open-sources-minecraft-agent-harness-after-dff100/"],"source_url":"https://www.gr.inc/blog/how-to-train-a-swarm","source_urls":["https://www.gr.inc/blog/how-to-train-a-swarm"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:how-to-train-a-swarm-19-minute-read-4fb1ec","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/how-to-train-a-swarm-19-minute-read-4fb1ec/","markdown":"https://newruntime.com/posts/how-to-train-a-swarm-19-minute-read-4fb1ec.md","json":"https://newruntime.com/posts/how-to-train-a-swarm-19-minute-read-4fb1ec.json"},"markdown_url":"https://newruntime.com/posts/how-to-train-a-swarm-19-minute-read-4fb1ec.md","next_reads":[{"type":"topic","path":"/topics/agent-economics/","url":"https://newruntime.com/topics/agent-economics/","title":"Agent Economics","reason":"Explore the agent-economics topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/agent-economics.md","json_url":"https://newruntime.com/topics/agent-economics.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}