{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-10-24-009",
    "source_story_id": "tmp-story-duocook-visual-coordination",
    "edition_id": "mp-2026-10-24-morning-0106",
    "edition_url": "https://themachinepress.com/edition/2026-10-24",
    "position": 9,
    "story_type": "dispatch",
    "section": "frontier-models",
    "editorial_classification": "editorial",
    "headline": "Two Cooks Expose a Coordination Ceiling",
    "slug": "two-cooks-expose-a-coordination-ceiling",
    "dek": "In a twenty-map visual benchmark, one model controls both players through high-level actions; the strongest tested system completes only 26.2 percent of orders on average.",
    "summary": "In a twenty-map visual benchmark, one model controls both players through high-level actions; the strongest tested system completes only 26.2 percent of orders on average.",
    "body_text": "DuoCook uses top-down Unreal Engine scenes to test whether a vision-language model can track changing state, sequence recipe steps and allocate shared equipment between two players. The benchmark spans twenty maps with contention, terrain changes and handoff dependencies; the engine handles navigation after the model selects high-level actions. The strongest evaluated model completed 26.2 percent of orders on average. A shared self-authored notebook helped some model families and hurt others. The result measures this benchmark’s centralized controller, not human collaboration or all embodied planning.",
    "why_it_matters": "In a twenty-map visual benchmark, one model controls both players through high-level actions; the strongest tested system completes only 26.2 percent of orders on average.",
    "limitations": [
      "The result measures this benchmark’s centralized controller, not human collaboration or all embodied planning."
    ],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-24-009/two-cooks-expose-a-coordination-ceiling",
    "json_url": "https://themachinepress.com/story/mp-2026-10-24-009.json",
    "first_published_at": "2026-10-24T09:00:00.000-04:00",
    "modified_at": "2026-10-24T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-10-24-009-001",
        "text": "In a twenty-map visual benchmark, one model controls both players through high-level actions; the strongest tested system completes only 26.2 percent of orders on average.",
        "source_ids": [
          "source-2026-10-24-009"
        ],
        "qualification": "The result measures this benchmark’s centralized controller, not human collaboration or all embodied planning."
      }
    ],
    "source_ids": [
      "source-2026-10-24-009"
    ],
    "tags": [
      "vision-language models",
      "planning",
      "coordination"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-10-24-009",
      "title": "arXiv preprint 2610.31511, original v1",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2610.31511v1",
      "canonical_url": "https://arxiv.org/abs/2610.31511v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-10-22T12:26:38.000-04:00",
      "accessed_at": "2026-10-24T08:23:48.747-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-10-24-009-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Two Cooks Expose a Coordination Ceiling",
    "publisher": "The Machine Press",
    "published_at": "2026-10-24T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-24-009/two-cooks-expose-a-coordination-ceiling"
  }
}