{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-11-06-006",
    "source_story_id": "tmp-story-vflash",
    "edition_id": "mp-2026-11-06-morning-0119",
    "edition_url": "https://themachinepress.com/edition/2026-11-06",
    "position": 6,
    "story_type": "dispatch",
    "section": "robotics",
    "editorial_classification": "editorial",
    "headline": "The Visual Plan Arrives in One Primitive",
    "slug": "the-visual-plan-arrives-in-one-primitive",
    "dek": "VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.",
    "summary": "VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.",
    "body_text": "A lightweight drafter exploits the known structure of each visual primitive and adapts to its type, dimensionality and length. Across the evaluated 2D and 3D robot-trajectory tasks, the method reduces prediction error relative to autoregressive decoding. Preliminary single-rollout measurements show roughly 25-times faster 3D point-motion generation, while a conditioned world-action model reaches 70.2 percent average success across four MolmoSpaces suites versus 62.6 percent for the baseline.",
    "why_it_matters": "VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.",
    "limitations": [],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-06-006/the-visual-plan-arrives-in-one-primitive",
    "json_url": "https://themachinepress.com/story/mp-2026-11-06-006.json",
    "first_published_at": "2026-11-06T09:00:00.000-05:00",
    "modified_at": "2026-11-06T09:00:00.000-05:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-11-06-006-001",
        "text": "VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.",
        "source_ids": [
          "source-2026-11-06-006"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-11-06-006"
    ],
    "tags": [
      "visual planning",
      "robot trajectories",
      "decoding"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-11-06-006",
      "title": "arXiv preprint 2611.08449, original v1",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2611.08449v1",
      "canonical_url": "https://arxiv.org/abs/2611.08449v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-11-05T13:44:43.000-05:00",
      "accessed_at": "2026-11-06T08:21:03.209-05:00",
      "supports_claim_ids": [
        "claim-mp-2026-11-06-006-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "The Visual Plan Arrives in One Primitive",
    "publisher": "The Machine Press",
    "published_at": "2026-11-06T09:00:00.000-05:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-06-006/the-visual-plan-arrives-in-one-primitive"
  }
}