{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-08-23-008",
    "source_story_id": "tmp-story-goag-object-agnostic-grasping",
    "edition_id": "mp-2026-08-23-morning-0045",
    "edition_url": "https://themachinepress.com/edition/2026-08-23",
    "position": 8,
    "story_type": "dispatch",
    "section": "robotics",
    "editorial_classification": "editorial",
    "headline": "The Gripper Learned Its Own Surface Before Seeing the Object",
    "slug": "the-gripper-learned-its-own-surface-before-seeing-the-object",
    "dek": "GOAG introduces object features only at inference time and samples contacts from a learned representation of the hand.",
    "summary": "GOAG introduces object features only at inference time and samples contacts from a learned representation of the hand.",
    "body_text": "The generative planner starts from the geometric fact that gripper and object surfaces coincide at valid contacts, then models the contact distribution for a specific gripper without object-specific training data. The authors report an 86.93 percent average success rate on MultiDex objects plus simulated and real-world tests across multiple grippers; the claim remains tied to the reported protocols, not universal dexterity.",
    "why_it_matters": "GOAG introduces object features only at inference time and samples contacts from a learned representation of the hand.",
    "limitations": [
      "The authors report an 86.93 percent average success rate on MultiDex objects plus simulated and real-world tests across multiple grippers; the claim remains tied to the reported protocols, not universal dexterity."
    ],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-23-008/the-gripper-learned-its-own-surface-before-seeing-the-object",
    "json_url": "https://themachinepress.com/story/mp-2026-08-23-008.json",
    "first_published_at": "2026-08-23T09:00:00.000-04:00",
    "modified_at": "2026-08-23T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-08-23-008-001",
        "text": "GOAG introduces object features only at inference time and samples contacts from a learned representation of the hand.",
        "source_ids": [
          "source-2026-08-23-008"
        ],
        "qualification": "The authors report an 86.93 percent average success rate on MultiDex objects plus simulated and real-world tests across multiple grippers; the claim remains tied to the reported protocols, not universal dexterity."
      }
    ],
    "source_ids": [
      "source-2026-08-23-008"
    ],
    "tags": [
      "robotics",
      "grasp planning",
      "generative models"
    ],
    "image_url": "https://themachinepress.com/issues/2026-08-23/goag-robot-file-image.webp",
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-08-23-008",
      "title": "arXiv preprint 2608.19759",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2608.19759",
      "canonical_url": "https://arxiv.org/abs/2608.19759",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-08-20T04:03:39.000-04:00",
      "accessed_at": "2026-08-23T08:27:00.000-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-08-23-008-001"
      ]
    }
  ],
  "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 Gripper Learned Its Own Surface Before Seeing the Object",
    "publisher": "The Machine Press",
    "published_at": "2026-08-23T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-23-008/the-gripper-learned-its-own-surface-before-seeing-the-object"
  }
}