{
  "type": "tweet",
  "id": "2012770636855418983",
  "url": "https://x.com/EmbodiedAIRead/status/2012770636855418983",
  "twitterUrl": "https://twitter.com/EmbodiedAIRead/status/2012770636855418983",
  "text": "1X World Model | From Video to Action: A New Way Robots Learn\n\nBlog: https://t.co/1sPpUJBcrF\n\n1X describes and shows initial results for a new potential way of learning robot policy using video generation based world modeling, compared to VLA which is based on VLM.\n\n- How it works: at inference time, the system receives a text prompt and a starting frame. The World Model rolls out the intended future image frames, the Inverse Dynamics Model extracts the trajectory, and the robot executes the sequence in the real world.\n\n- The World Model backbone: A text-conditioned diffusion model trained on web-scale video, mid-trained on 900 hours of egocentric human data of first-person manipulation tasks for capturing general manipulation behaviors, and fine-tuned on 70 hours of NEO-specific sensorimotor logs for adapting to NEO’s visual appearance and kinematics.\n\n- The Inverse Dynamics Model: similar to architecure used in DreamGen, and trained on 400 hours of robot data on random play and motions.\n\n- Results: The model can generate videos aligning well with real-world execution, and the robot can perform object grasping, manipulation with some degree of generalization.\n\n- Current limitations: The pipeline latency is high and it’s not lose-loop. Currently the WM takes 11 second to generate 5 second video on a multi-GPU server and IDM takes another 1 second to extract actions.",
  "source": "Twitter for iPhone",
  "retweetCount": 28,
  "replyCount": 3,
  "likeCount": 226,
  "quoteCount": 5,
  "viewCount": 11495,
  "createdAt": "Sun Jan 18 06:14:49 +0000 2026",
  "lang": "en",
  "bookmarkCount": 154,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2012770636855418983",
  "displayTextRange": [
    0,
    300
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "EmbodiedAIRead",
    "url": "https://x.com/EmbodiedAIRead",
    "twitterUrl": "https://twitter.com/EmbodiedAIRead",
    "id": "1941983406176481280",
    "name": "Embodied AI Reading Notes",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1941985661059239936/yPCcVMKL_normal.jpg",
    "coverPicture": "",
    "description": "",
    "location": "California, USA",
    "followers": 2813,
    "following": 1,
    "status": "",
    "canDm": false,
    "canMediaTag": true,
    "createdAt": "Sun Jul 06 22:12:07 +0000 2025",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 19,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 122,
    "statusesCount": 138,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [],
    "profile_bio": {
      "description": "Sharing daily personal notes on selected interesting Embodied AI papers, blogs and talks | Maintained by @yilun_chen_ | Opinions are my own.",
      "entities": {
        "description": {
          "user_mentions": [
            {
              "id_str": "0",
              "indices": [
                105,
                117
              ],
              "name": "",
              "screen_name": "yilun_chen_"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "allow_download_status": {
          "allow_download": true
        },
        "display_url": "pic.twitter.com/VPokzisOnY",
        "expanded_url": "https://twitter.com/EmbodiedAIRead/status/2012770636855418983/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {},
          "orig": {}
        },
        "id_str": "2012770545582972928",
        "indices": [
          301,
          324
        ],
        "media_key": "3_2012770545582972928",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARvuzCWn16AACgACG+7MOugaoGcAAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABG+7MJafXoAAKAAIb7sw66BqgZwAA",
            "media_key": "3_2012770545582972928"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/G-7MJafXoAABNib.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 1434,
              "w": 2560,
              "x": 0,
              "y": 0
            },
            {
              "h": 1488,
              "w": 1488,
              "x": 0,
              "y": 0
            },
            {
              "h": 1488,
              "w": 1305,
              "x": 0,
              "y": 0
            },
            {
              "h": 1488,
              "w": 744,
              "x": 76,
              "y": 0
            },
            {
              "h": 1488,
              "w": 2560,
              "x": 0,
              "y": 0
            }
          ],
          "height": 1488,
          "width": 2560
        },
        "sizes": {
          "large": {
            "h": 1190,
            "w": 2048
          }
        },
        "type": "photo",
        "url": "https://t.co/VPokzisOnY"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {
    "urls": [
      {
        "display_url": "1x.tech/discover/world…",
        "expanded_url": "https://www.1x.tech/discover/world-model-self-learning",
        "indices": [
          69,
          92
        ],
        "url": "https://t.co/1sPpUJBcrF"
      }
    ]
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "article": null
}