{ "schema_version": "paper_public_manifest_v1", "paper_id": "liangwenfeng-world-model_2026_07", "slug": "liangwenfeng-world-model", "title": "This statement is missing a premise: whether your data pipeline is connected", "authors": [], "source": { "arxiv_id": "", "pdf_url": "", "project_url": "https://mp.weixin.qq.com/s/AWsSjcT9NYbj1W8SWXgb_w", "github_url": "", "huggingface_url": "", "original_source": "https://mp.weixin.qq.com/s/AWsSjcT9NYbj1W8SWXgb_w" }, "site": { "post_url": "/posts/liangwenfeng-world-model", "canonical_url": "https://haiguangboy.com/posts/liangwenfeng-world-model", "cover_image": "https://static.haiguangboy.com/papers/liangwenfeng-world-model/cover.webp" }, "taxonomy": { "domain": "embodied_ai", "track": "world_model", "tasks": [ "embodied_ai", "world_model", "vla", "action_generation", "robotics", "Embodied Intelligence", "World Models", "deepseek", "Technical Approach" ], "related_topics": [ { "paper_id": "wx_界面新闻_20260605_2026_06", "title": "wx_Interface News_20260605", "url": "https://haiguangboy.com/posts/wx_界面新闻_20260605", "relation": "contrast", "summary": "★Liang Wenfeng's AGI ladder: CoT→Agent→Continuous Learning→Gradual Singularity→Embodied Intelligence, explicitly stating that 'the end point of intelligence may all be embodied' contradicts the route bet: building the brain, not the body—betting on the identity of a foundation model company and rejecting the track of motor skills", "strength": "strong" }, { "paper_id": "latent_action_pretraining_through_world_modeling_2026_07", "title": "LAWM: Why Action Labels Become a Burden", "url": "https://haiguangboy.com/posts/latent_action_pretraining_through_world_modeling", "relation": "same_track", "summary": "LAWM: Why Action Labels Become a Burden", "strength": "strong" }, { "paper_id": "latepost_xuhuazhe_202603_2026_03", "title": "latepost_xuhuazhe_202603", "url": "https://haiguangboy.com/posts/latepost_xuhuazhe_202603", "relation": "same_track", "summary": "★Liang Wenfeng's AGI ladder: CoT→Agent→Continuous Learning→Gradual Singularity→Embodied Intelligence, explicitly stating that 'the end point of intelligence may all be embodied' validates the three entrepreneurial signals: RL single-point tasks solved + high scaling law ceiling + reusable Agent framework", "strength": "strong" }, { "paper_id": "orca_2026_07", "title": "π0.5 trembles in place when failing to grab a spoon, but Orca goes further with physical intuition learned from watching videos", "url": "https://haiguangboy.com/posts/orca", "relation": "same_track", "summary": "The key to a world model is a readable state", "strength": "strong" }, { "paper_id": "wx_中国财富_20260719_2026_07", "title": "Qianxun Intelligence Han Fengtao: Embodied Intelligence Enters the 'Speed Competition' Stage", "url": "https://haiguangboy.com/posts/qianxun-general-brain-data-loop", "relation": "same_track", "summary": "★Continuous learning is a common gap between LLMs and embodied intelligence, but embodied companies are giving timelines for general capabilities while continuous learning remains unsolved validates ★Time point prediction: Embodied Intelligence's 'ChatGPT moment' is between late 2027 and early 2028—highly convergent with predictions from two other companies in the library", "strength": "strong" }, { "paper_id": "wx_晚点latepost_20260718_2026_07", "title": "Sudu Technology's WAIC Debut: Reality is the Endpoint for Robots, Not the Boundary of Training", "url": "https://haiguangboy.com/posts/sudu-third-position", "relation": "same_track", "summary": "★★Core judgment: The implicit premise of 'world models are unrelated to the ceiling of intelligence' is that 'the data pipeline is connected'—LLMs are connected, so it's a detour; robots are not, so it's a bridge validates the precise formulation of the simulation-first training route: almost no real-machine data in pretraining, but failure retries after deployment plus a small amount of real-machine RL later", "strength": "medium" } ] }, "ruling": { "importance_score": 3.0, "one_sentence": "This statement is missing a premise: whether your data pipeline is connected" }, "asset_base_url": "https://static.haiguangboy.com/papers/liangwenfeng-world-model", "assets": [ { "type": "public_brief", "object_key": "papers/liangwenfeng-world-model/public_brief.md", "content_type": "text/markdown; charset=utf-8", "upload_status": "uploaded", "bucket": "paper-assets", "url": "https://static.haiguangboy.com/papers/liangwenfeng-world-model/public_brief.md", "role": "public_brief", "size_bytes": 5042 }, { "type": "public_manifest", "object_key": "papers/liangwenfeng-world-model/public_manifest.json", "content_type": "application/json; charset=utf-8", "upload_status": "uploaded", "bucket": "paper-assets", "url": "https://static.haiguangboy.com/papers/liangwenfeng-world-model/public_manifest.json", "role": "public_manifest", "size_bytes": 5153 } ], "published_at": "2026-07-23T15:00:13+08:00", "created_at": "2026-07-23T15:00:13+08:00", "updated_at": "2026-09-02T10:55:06+08:00", "analyst_take": { "type": "author_opinion", "text": "💡The real disagreement lies elsewhere\n\nIn the comparison, the only strong contradiction edge is not about world models—it's Liang Wenfeng's 'the end point of intelligence is embodied' colliding with the independent variable robot's bet on 'doing the brain, not the body.' **The real question is not whether to have world models, but whether you can do only the brain.**\n\nThere's another unresolved one: LeCun believes the entire LLM path is wrong, and the JEPA-style world model **is** the path to intelligence. That is a genuine disagreement that definitional clarification cannot resolve—the verdict may come even later than 2028." } }