{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-10-11-009",
    "source_story_id": "tmp-story-worldalign-decoupled-reward",
    "edition_id": "mp-2026-10-11-morning-0093",
    "edition_url": "https://themachinepress.com/edition/2026-10-11",
    "position": 9,
    "story_type": "dispatch",
    "section": "frontier-models",
    "editorial_classification": "editorial",
    "headline": "Static Geometry and Moving Subjects Stopped Sharing One Judge",
    "slug": "static-geometry-and-moving-subjects-stopped-sharing-one-judge",
    "dek": "WorldAlign scores background structure and subject motion with separate priors before post-training a video generator.",
    "summary": "WorldAlign scores background structure and subject motion with separate priors before post-training a video generator.",
    "body_text": "WorldAlign separates two kinds of 4D consistency that a single reward can blur. Static regions are checked through semantically masked geometric reprojection, with an additional camera-motion reward to discourage trivial stillness. Dynamic subjects are judged by sample-specific checklists covering motion, physical plausibility, shape and texture. Online post-training with these decoupled rewards improved both static and dynamic consistency across two image-to-video generators without suppressing reported motion. The dynamic component still relies on a vision-language model as judge, so the result is evidence for a better reward design rather than a direct measure of physical truth.",
    "why_it_matters": "WorldAlign scores background structure and subject motion with separate priors before post-training a video generator.",
    "limitations": [],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-11-009/static-geometry-and-moving-subjects-stopped-sharing-one-judge",
    "json_url": "https://themachinepress.com/story/mp-2026-10-11-009.json",
    "first_published_at": "2026-10-11T09:00:00.000-04:00",
    "modified_at": "2026-10-11T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-10-11-009-001",
        "text": "WorldAlign scores background structure and subject motion with separate priors before post-training a video generator.",
        "source_ids": [
          "source-2026-10-11-009"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-10-11-009"
    ],
    "tags": [
      "video generation",
      "4D consistency",
      "reward modeling"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-10-11-009",
      "title": "arXiv preprint 2610.12382, original v1",
      "publisher": "arXiv / submitting research authors",
      "url": "https://arxiv.org/abs/2610.12382v1",
      "canonical_url": "https://arxiv.org/abs/2610.12382v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-10-08T13:35:19.000-04:00",
      "accessed_at": "2026-10-11T08:21:27.764-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-10-11-009-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Static Geometry and Moving Subjects Stopped Sharing One Judge",
    "publisher": "The Machine Press",
    "published_at": "2026-10-11T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-11-009/static-geometry-and-moving-subjects-stopped-sharing-one-judge"
  }
}