{
  "blyg": "0.3",
  "id": "45nswzb123a15dv3a48dpy4nwg",
  "kind": "fragment",
  "origin": "https://www.msweet.net/notes/",
  "page": "875-violet-shrews",
  "author": {
    "name": "Matthew McDowell-Sweet",
    "url": "https://www.msweet.net/"
  },
  "created": "2026-11-14T13:57:10Z",
  "updated": "2026-11-14T13:57:10Z",
  "version": 1,
  "content_md": "New S3 just dropped.\n\n> Can AI agents learn to work together without being told how? We train swarms of language models with no leader, no roles, and no turn order. They share one reward, message each other whenever they want, and learn to organize during training.\n\n[s³: Simple Scaling of Swarms via Cooperative Multi-agent RL ↗](https://scalingswarms.com/)",
  "content_html": "<p>New S3 just dropped.</p>\n<blockquote><p>Can AI agents learn to work together without being told how? We train swarms of language models with no leader, no roles, and no turn order. They share one reward, message each other whenever they want, and learn to organize during training.</p></blockquote>\n<p class=\"source\"><a href=\"https://scalingswarms.com/\">s³: Simple Scaling of Swarms via Cooperative Multi-agent RL ↗</a></p>",
  "content_hash": "sha256:15cb51cb0ed2efa4db509ac9fbe536fde4dc150a0a87e0a03cac71a41e90ff41",
  "media": [],
  "changelog": [
    {
      "version": 1,
      "at": "2026-11-14T13:57:10Z",
      "note": null
    }
  ]
}
