{"id":"442c4f8aa1f089b723b44eb9c7b09199ea5d584c5ec94e3e380fa85dd0d005a4","title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","argument":"The authors of TAPESTRY describe a way to use a pretrained video diffusion transformer for robot control without generating video at inference. They represent robot actions as tokens in the video model’s latent space, allowing video and action information to use the same transformer backbone. The model is trained with video diffusion, joint video+action diffusion, and action-only diffusion objectives. At test time, it generates action sequences in a single DDIM step, which the authors say produces smooth action chunks at 84 ms per chunk on consumer hardware. They report strong performance in simulation benchmarks and real-robot experiments, with lower inference cost than comparable joint video–action policies. If the reported performance and speed hold in deployment, this approach could reduce the cost of using video-pretrained models for robot control; the abstract does not establish performance across other tasks or settings. What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The abstract does not name the simulation benchmarks or real-robot tasks, or provide specific performance results. The abstract reports inference speed on consumer hardware but does not identify the hardware or provide a comparison of specific inference costs. The reported results do not establish performance across tasks or deployment settings beyond those evaluated.","speaker":{"name":"Collins, Jeremy A., Lee, Seungjae, Wachakorn, Rachanon, Srinivasan, Krishnan, Garg, Animesh, Guizilini, Vitor Campagnolo, Shah, Paarth","roleAtSource":"Article author"},"source":{"title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2610.36823","publishedAt":"2026-10-27","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors introduce TAPESTRY, which repurposes a pretrained video diffusion transformer as a robotic policy without video generation at inference.","excerpt":"repurposes a pretrained video diffusion transformer into a robotic policy without requiring video generation at inference","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"},{"text":"TAPESTRY represents robot actions as single-patch tokens in the video model’s latent space and processes them with the same transformer backbone used for video.","excerpt":"treat robot actions as single-patch tokens in the video model's latent space","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"},{"text":"At test time, the model generates action tokens in one DDIM step; the authors report 84 ms per action chunk on consumer hardware.","excerpt":"generates a compact action token sequence in a single DDIM step from pure noise, producing smooth action chunks at 84 ms per action chunk on consumer hardware","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"},{"text":"The authors report strong performance across simulation benchmarks and real-robot experiments, with substantially lower inference cost than comparable joint video–action policies.","excerpt":"Across simulation benchmarks and real-robot experiments, TAPESTRY achieves strong performance with substantially lower inference cost than comparable joint video--action policies.","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36823","checkedAt":"2026-11-07T02:08:07.701Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","summary":"The authors of TAPESTRY describe a way to use a pretrained video diffusion transformer for robot control without generating video at inference. They represent robot actions as tokens in the video model’s latent space, allowing video and action information to use the same transformer backbone. The model is trained with video diffusion, joint video+action diffusion, and action-only diffusion objectives. At test time, it generates action sequences in a single DDIM step, which the authors say produces smooth action chunks at 84 ms per chunk on consumer hardware. They report strong performance in simulation benchmarks and real-robot experiments, with lower inference cost than comparable joint video–action policies. If the reported performance and speed hold in deployment, this approach could reduce the cost of using video-pretrained models for robot control; the abstract does not establish performance across other tasks or settings. What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?","question":"What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?","claims":[{"text":"The authors introduce TAPESTRY, which repurposes a pretrained video diffusion transformer as a robotic policy without video generation at inference.","excerpt":"repurposes a pretrained video diffusion transformer into a robotic policy without requiring video generation at inference","support":"source-stated"},{"text":"TAPESTRY represents robot actions as single-patch tokens in the video model’s latent space and processes them with the same transformer backbone used for video.","excerpt":"treat robot actions as single-patch tokens in the video model's latent space","support":"source-stated"},{"text":"At test time, the model generates action tokens in one DDIM step; the authors report 84 ms per action chunk on consumer hardware.","excerpt":"generates a compact action token sequence in a single DDIM step from pure noise, producing smooth action chunks at 84 ms per action chunk on consumer hardware","support":"source-stated"},{"text":"The authors report strong performance across simulation benchmarks and real-robot experiments, with substantially lower inference cost than comparable joint video–action policies.","excerpt":"Across simulation benchmarks and real-robot experiments, TAPESTRY achieves strong performance with substantially lower inference cost than comparable joint video--action policies.","support":"source-stated"}],"limitations":["The abstract does not name the simulation benchmarks or real-robot tasks, or provide specific performance results.","The abstract reports inference speed on consumer hardware but does not identify the hardware or provide a comparison of specific inference costs.","The reported results do not establish performance across tasks or deployment settings beyond those evaluated."],"mediaId":"b69a9401699cde33eea8e73f7fad95897bb2716973fda600729e1fed74bc19ec","correctionReason":null,"id":"442c4f8aa1f089b723b44eb9c7b09199ea5d584c5ec94e3e380fa85dd0d005a4","publicationId":"0ac2fe74-2ea9-4afd-837c-d5c1cae8a21f","version":1,"sourceUrl":"https://arxiv.org/abs/2610.36823","sourceTitle":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-10-27","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-10-28T05:48:36.212Z","createdAt":"2026-11-07T00:01:31.436Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"b69a9401699cde33eea8e73f7fad95897bb2716973fda600729e1fed74bc19ec","kind":"article","url":"https://arxiv.org/abs/2610.36823","sourceUrl":"https://arxiv.org/abs/2610.36823","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-10-29T23:32:48.938Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36823","checkedAt":"2026-11-07T02:08:07.701Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"b69a9401699cde33eea8e73f7fad95897bb2716973fda600729e1fed74bc19ec","confidence":0.94,"provider":"typesafe","model":"jev-1.13.0","inputHash":"94e500faa00a72c916c0cde87c0f7eb5919e8bb0af80e9572eec738ea3db24f7","decidedAt":"2026-10-29T23:32:49.512Z"},"provenance":{"textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observationId":"b1a7660b1759036b51d4507fa026d87300de50a627d7f803d9eb0a0029890637","retrievedAt":"2026-10-29T23:32:48.837Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"5f2ff771-8f89-44f4-8ec3-b7231bd33dd0"},"reviewer":{"model":"gpt-6-luna","runId":"ab69c42a-24fd-408d-888a-002a57204349"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-07T00:01:31.436Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Collins, Jeremy A.","url":null,"kind":"unknown"},{"name":"Lee, Seungjae","url":null,"kind":"unknown"},{"name":"Wachakorn, Rachanon","url":null,"kind":"unknown"},{"name":"Srinivasan, Krishnan","url":null,"kind":"unknown"},{"name":"Garg, Animesh","url":null,"kind":"unknown"},{"name":"Guizilini, Vitor Campagnolo","url":null,"kind":"unknown"},{"name":"Shah, Paarth","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Collins, Jeremy A.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Collins, Jeremy A.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Lee, Seungjae","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Lee, Seungjae","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Wachakorn, Rachanon","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wachakorn, Rachanon","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Srinivasan, Krishnan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Srinivasan, Krishnan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Garg, Animesh","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Garg, Animesh","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Guizilini, Vitor Campagnolo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Guizilini, Vitor Campagnolo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Shah, Paarth","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shah, Paarth","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}}],"participationCoverage":{"total":7,"returned":7,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2610.36823","snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z","basis":"publisher-page-metadata","sourceLinks":49,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/442c4f8aa1f089b723b44eb9c7b09199ea5d584c5ec94e3e380fa85dd0d005a4/source"}},"canonicalUrl":"https://escapemeridian.com/readings/tapestry-a-single-backbone-for-video-generation-and-robot-442c4f8a","formats":{"html":"/readings/tapestry-a-single-backbone-for-video-generation-and-robot-442c4f8a","markdown":"/readings/tapestry-a-single-backbone-for-video-generation-and-robot-442c4f8a.md","json":"/readings/tapestry-a-single-backbone-for-video-generation-and-robot-442c4f8a.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-pre-training","name":"Pre-training","explanation":"A model's initial training on a large dataset, before further adaptation for particular uses.","referenceUrl":"https://developers.google.com/machine-learning/glossary#pre-training"},{"id":"term-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"},{"id":"term-inference","name":"Model inference","explanation":"Using a trained model to produce a prediction or response to new input.","referenceUrl":"https://developers.google.com/machine-learning/glossary#inference"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"},{"id":"time","name":"TIME","kind":"organization","context":"TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.","referenceUrls":["https://time.com/"],"path":"/entities/time","json":"/entities/time.json","markdown":"/entities/time.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"442c4f8aa1f089b723b44eb9c7b09199ea5d584c5ec94e3e380fa85dd0d005a4","revision":"0ac2fe74-2ea9-4afd-837c-d5c1cae8a21f","revisionKind":"publication-id","title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","url":"https://escapemeridian.com/readings/tapestry-a-single-backbone-for-video-generation-and-robot-442c4f8a"},"source":{"type":"source","id":"https://arxiv.org/abs/2610.36823","revision":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","revisionKind":"sha256-source-text","title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","url":"https://arxiv.org/abs/2610.36823"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"442c4f8aa1f089b723b44eb9c7b09199ea5d584c5ec94e3e380fa85dd0d005a4","title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","argument":"The authors of TAPESTRY describe a way to use a pretrained video diffusion transformer for robot control without generating video at inference. They represent robot actions as tokens in the video model’s latent space, allowing video and action information to use the same transformer backbone. The model is trained with video diffusion, joint video+action diffusion, and action-only diffusion objectives. At test time, it generates action sequences in a single DDIM step, which the authors say produces smooth action chunks at 84 ms per chunk on consumer hardware. They report strong performance in simulation benchmarks and real-robot experiments, with lower inference cost than comparable joint video–action policies. If the reported performance and speed hold in deployment, this approach could reduce the cost of using video-pretrained models for robot control; the abstract does not establish performance across other tasks or settings. What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The abstract does not name the simulation benchmarks or real-robot tasks, or provide specific performance results. The abstract reports inference speed on consumer hardware but does not identify the hardware or provide a comparison of specific inference costs. The reported results do not establish performance across tasks or deployment settings beyond those evaluated.","speaker":{"name":"Collins, Jeremy A., Lee, Seungjae, Wachakorn, Rachanon, Srinivasan, Krishnan, Garg, Animesh, Guizilini, Vitor Campagnolo, Shah, Paarth","roleAtSource":"Article author"},"source":{"title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2610.36823","publishedAt":"2026-10-27","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors introduce TAPESTRY, which repurposes a pretrained video diffusion transformer as a robotic policy without video generation at inference.","excerpt":"repurposes a pretrained video diffusion transformer into a robotic policy without requiring video generation at inference","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"},{"text":"TAPESTRY represents robot actions as single-patch tokens in the video model’s latent space and processes them with the same transformer backbone used for video.","excerpt":"treat robot actions as single-patch tokens in the video model's latent space","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"},{"text":"At test time, the model generates action tokens in one DDIM step; the authors report 84 ms per action chunk on consumer hardware.","excerpt":"generates a compact action token sequence in a single DDIM step from pure noise, producing smooth action chunks at 84 ms per action chunk on consumer hardware","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"},{"text":"The authors report strong performance across simulation benchmarks and real-robot experiments, with substantially lower inference cost than comparable joint video–action policies.","excerpt":"Across simulation benchmarks and real-robot experiments, TAPESTRY achieves strong performance with substantially lower inference cost than comparable joint video--action policies.","url":"https://arxiv.org/abs/2610.36823","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36823","checkedAt":"2026-11-07T02:08:07.701Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","summary":"The authors of TAPESTRY describe a way to use a pretrained video diffusion transformer for robot control without generating video at inference. They represent robot actions as tokens in the video model’s latent space, allowing video and action information to use the same transformer backbone. The model is trained with video diffusion, joint video+action diffusion, and action-only diffusion objectives. At test time, it generates action sequences in a single DDIM step, which the authors say produces smooth action chunks at 84 ms per chunk on consumer hardware. They report strong performance in simulation benchmarks and real-robot experiments, with lower inference cost than comparable joint video–action policies. If the reported performance and speed hold in deployment, this approach could reduce the cost of using video-pretrained models for robot control; the abstract does not establish performance across other tasks or settings. What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?","question":"What performance and inference costs does TAPESTRY achieve on specific benchmarks and real-robot tasks?","claims":[{"text":"The authors introduce TAPESTRY, which repurposes a pretrained video diffusion transformer as a robotic policy without video generation at inference.","excerpt":"repurposes a pretrained video diffusion transformer into a robotic policy without requiring video generation at inference","support":"source-stated"},{"text":"TAPESTRY represents robot actions as single-patch tokens in the video model’s latent space and processes them with the same transformer backbone used for video.","excerpt":"treat robot actions as single-patch tokens in the video model's latent space","support":"source-stated"},{"text":"At test time, the model generates action tokens in one DDIM step; the authors report 84 ms per action chunk on consumer hardware.","excerpt":"generates a compact action token sequence in a single DDIM step from pure noise, producing smooth action chunks at 84 ms per action chunk on consumer hardware","support":"source-stated"},{"text":"The authors report strong performance across simulation benchmarks and real-robot experiments, with substantially lower inference cost than comparable joint video–action policies.","excerpt":"Across simulation benchmarks and real-robot experiments, TAPESTRY achieves strong performance with substantially lower inference cost than comparable joint video--action policies.","support":"source-stated"}],"limitations":["The abstract does not name the simulation benchmarks or real-robot tasks, or provide specific performance results.","The abstract reports inference speed on consumer hardware but does not identify the hardware or provide a comparison of specific inference costs.","The reported results do not establish performance across tasks or deployment settings beyond those evaluated."],"mediaId":"b69a9401699cde33eea8e73f7fad95897bb2716973fda600729e1fed74bc19ec","correctionReason":null,"id":"442c4f8aa1f089b723b44eb9c7b09199ea5d584c5ec94e3e380fa85dd0d005a4","publicationId":"0ac2fe74-2ea9-4afd-837c-d5c1cae8a21f","version":1,"sourceUrl":"https://arxiv.org/abs/2610.36823","sourceTitle":"TAPESTRY: A Single Backbone for Video Generation and Robot Control","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-10-27","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-10-28T05:48:36.212Z","createdAt":"2026-11-07T00:01:31.436Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"b69a9401699cde33eea8e73f7fad95897bb2716973fda600729e1fed74bc19ec","kind":"article","url":"https://arxiv.org/abs/2610.36823","sourceUrl":"https://arxiv.org/abs/2610.36823","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-10-29T23:32:48.938Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2610.36823","checkedAt":"2026-11-07T02:08:07.701Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"b69a9401699cde33eea8e73f7fad95897bb2716973fda600729e1fed74bc19ec","confidence":0.94,"provider":"typesafe","model":"jev-1.13.0","inputHash":"94e500faa00a72c916c0cde87c0f7eb5919e8bb0af80e9572eec738ea3db24f7","decidedAt":"2026-10-29T23:32:49.512Z"},"provenance":{"textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observationId":"b1a7660b1759036b51d4507fa026d87300de50a627d7f803d9eb0a0029890637","retrievedAt":"2026-10-29T23:32:48.837Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"5f2ff771-8f89-44f4-8ec3-b7231bd33dd0"},"reviewer":{"model":"gpt-6-luna","runId":"ab69c42a-24fd-408d-888a-002a57204349"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-07T00:01:31.436Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Collins, Jeremy A.","url":null,"kind":"unknown"},{"name":"Lee, Seungjae","url":null,"kind":"unknown"},{"name":"Wachakorn, Rachanon","url":null,"kind":"unknown"},{"name":"Srinivasan, Krishnan","url":null,"kind":"unknown"},{"name":"Garg, Animesh","url":null,"kind":"unknown"},{"name":"Guizilini, Vitor Campagnolo","url":null,"kind":"unknown"},{"name":"Shah, Paarth","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Collins, Jeremy A.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Collins, Jeremy A.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Lee, Seungjae","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Lee, Seungjae","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Wachakorn, Rachanon","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wachakorn, Rachanon","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Srinivasan, Krishnan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Srinivasan, Krishnan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Garg, Animesh","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Garg, Animesh","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Guizilini, Vitor Campagnolo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Guizilini, Vitor Campagnolo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}},{"actor":{"name":"Shah, Paarth","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shah, Paarth","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2610.36823","locator":"head/meta[name=author|citation_author]"},"assertionId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-07T02:08:07.701Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","sourceUrl":"https://arxiv.org/abs/2610.36823","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z"}}],"participationCoverage":{"total":7,"returned":7,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2610.36823","snapshotId":"43281a969561b27f459559b054d3b86fe70759251c4fb98f1cf60af81dfd2829","contentSha256":"9825aa493819c2de38f2c71413973e0610d469336dc8cc2536ff6cdddc704b78","textHash":"78484a3b2d5f4bcfd21c93ae84f031a5f74d692f44e820ba873452e019e64f88","observedAt":"2026-11-07T02:08:07.701Z","basis":"publisher-page-metadata","sourceLinks":49,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/442c4f8aa1f089b723b44eb9c7b09199ea5d584c5ec94e3e380fa85dd0d005a4/source"}}}],"audits":[{"publicationId":"0ac2fe74-2ea9-4afd-837c-d5c1cae8a21f","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading attributes the method, the speed figure and the performance claims to the paper's authors, using phrasing such as 'the authors say' and 'They report'. This matches the arXiv abstract and author list. No claim is presented as Escape Meridian's own finding.","excerpt":"We introduce Tokenizing Actions as Video Patches in a Shared Transformer Policy (TAPESTRY)"},{"dimension":"summary","status":"uncertain","explanation":"The mechanism is described faithfully, but the reading leaves out the abstract's motivation, so a non-specialist cannot see why a video model is useful for robots. The abstract says video models encode physical-interaction priors and that existing methods either generate video at test time, which is costly, or train action heads from scratch. Smallest repair: add one attributed sentence stating that motivation. Central terms such as diffusion transformer, latent space, DDIM and inference need term-library tooltips. The question is also repeated verbatim at the end of the summary.","excerpt":"Pretrained video generation models encode rich priors about physical interactions, but leveraging them for robot control typically requires either resource-intensive video generation at test time or action heads trained from scratch"},{"dimension":"date","status":"pass","explanation":"The originalPublished value, 2026-10-27, matches the v1 submission date on arXiv. The feed date of 2026-10-29 is later and correctly not used.","excerpt":"[Submitted on 27 Oct 2026]"},{"dimension":"support","status":"pass","explanation":"Each claim is backed by an exact abstract excerpt. The cost implication is explicitly conditional: 'If the reported performance and speed hold in deployment'. The limitations correctly note that the benchmarks are not named, the hardware is unspecified and the results are not quantified. The question follows from those gaps and does not lead the reader.","excerpt":"producing smooth action chunks at 84 ms per action chunk on consumer hardware and avoiding the cost of full video generation"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct link to the original arXiv abstract page, which matches the source URL. No other source media was found.","excerpt":null}],"createdAt":"2026-11-07T00:01:31.436Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}