{"as_of":"2026-08-07T22:05:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:9c798ea05b9dd52b29eef44b958179e72ab8e708839971b7a2718c122eadbcda","coverage":[{"denominator":86,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":86,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-06T23:49:21.427575Z","state":"measured"},{"denominator":86,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":86,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-07T06:34:17.273281+00:00","state":"measured"},{"denominator":0,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"cited_works","source_observed_at":null,"state":"measured"}],"external_citation_measurements":[],"inbound":[],"links":{"evidence":"/evidence","html":"/paper/2608.04458/citation-record","integrity":"/paper/2608.04458/integrity","json":"/paper/2608.04458/citation-record.json","paper":"/paper/2608.04458"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":"2601.03197","doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:22.741090Z","title":"Software-Defined Agentic Serving,","venue":null,"work_id":"44f3584c-e485-4e36-b345-d0ff8db3ff5c","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:12.261862Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:092ff2b2fe18476098fb01825f8ab10fd74632e70119c78adc72cc4f0f238b20","observation_id":"8d8d5178-6b93-4493-87da-f476a6ce2957","resolution":{"observed_at":"2026-08-06T23:49:22.828935Z","resolver_source":"raw_fallback","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:12.391886Z","title":"Micro-Sliced Virtual Processors to Hide the Effect of Discontinuous CPU Availability for Consolidated Systems,","venue":null,"work_id":null,"year":2014},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:12.391886Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:364af61f28dbe65911c076d9711024d90bf9defcfc2bbd32945871709942e7f4","observation_id":"c0c094a0-54c2-49b1-a376-7e293a6575be","resolution":{"observed_at":"2026-08-06T23:49:12.391886Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2312.11514","last_updated":"2024-07-30T23:37:20Z","snapshot_observed_at":"2026-07-06T17:04:50.608413Z","submitted_at":"2023-12-12T18:57:08Z","title":"LLM in a flash: Efficient Large Language Model Inference with Limited Memory","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2312.11514","snapshot_observed_at":"2026-08-06T23:49:12.501057Z","title":"LLM in a flash: Efficient Large Language Model Inference with Limited Memory,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:12.501057Z"},"links":{"cited_paper":"/paper/2312.11514","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:314c95272c65bcab844e9d6bc1bf9d40dfe26bdeed515ab98998183b2b11d41d","observation_id":"3cb3c67c-0732-4b33-b3fe-af6857632992","resolution":{"observed_at":"2026-08-06T23:49:12.501057Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:12.644868Z","title":"Agentic AI with AWS Databases,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:12.644868Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:fee4a29fc51852e8dca527febe965ce7df722b0a8eeb67c4db96611458a3ea4d","observation_id":"e799a95d-6bfa-4589-b24f-8a28811bd520","resolution":{"observed_at":"2026-08-06T23:49:12.644868Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:12.787123Z","title":"Agent Computers. Powering the Future of Agentic AI","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:12.787123Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:f186339a0a225fced439e7047c5f18f7f7411232ea5c929161714651a790a668","observation_id":"2dc3d2e9-a8fa-42ed-a424-143ed48e6d9b","resolution":{"observed_at":"2026-08-06T23:49:12.787123Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:12.948472Z","title":"Processing Architecture for Power Efficiency and Performance,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:12.948472Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:9a8368c7aef03f0316133d51bb33919aaf98818ca1a3c921cf13d3ea9f6c9f20","observation_id":"860d99e0-d274-41fe-8970-3db7d7de1ff6","resolution":{"observed_at":"2026-08-06T23:49:12.948472Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:13.254016Z","title":"Arm AGI CPU: The world’s most efficient agentic CPU,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.254016Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:10f79d329cc8c40f461dcce6ec08311edeb6bff66ca7ec9b70dd5310263a946b","observation_id":"961022b7-f037-4938-88bc-a9c026f6452d","resolution":{"observed_at":"2026-08-06T23:49:13.254016Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2507.19635","last_updated":"2025-07-25T19:02:42Z","snapshot_observed_at":"2026-08-06T14:16:12.509465Z","submitted_at":"2025-07-25T19:02:42Z","title":"Efficient and Scalable Agentic AI with Heterogeneous Systems","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2507.19635","snapshot_observed_at":"2026-08-06T23:49:13.400815Z","title":"Efficient and Scalable Agentic AI with Heterogeneous Systems,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.400815Z"},"links":{"cited_paper":"/paper/2507.19635","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:69e880682ec92f2700f1e8685df9c499661a0a1e46f77c5459f347bbd477ffe9","observation_id":"dd0f6d93-20a2-47bb-bf55-73ca6391e988","resolution":{"observed_at":"2026-08-06T23:49:13.400815Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2506.02153","last_updated":"2025-09-15T22:15:00Z","snapshot_observed_at":"2026-08-05T18:56:11.132981Z","submitted_at":"2025-06-02T18:35:16Z","title":"Small Language Models are the Future of Agentic AI","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2506.02153","snapshot_observed_at":"2026-08-06T23:49:13.524076Z","title":"Small Language Models are the Future of Agentic AI,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.524076Z"},"links":{"cited_paper":"/paper/2506.02153","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:4874fd021f56f78a8d53404b06638b616763285bc792dd4a40cc959f3402536b","observation_id":"6252d1da-b9a1-4228-825c-b1517bd9f5b9","resolution":{"observed_at":"2026-08-06T23:49:13.524076Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2604.03143","last_updated":"2026-04-03T16:04:40Z","snapshot_observed_at":"2026-07-06T22:52:25.307426Z","submitted_at":"2026-04-03T16:04:40Z","title":"TokenDance: Scaling Multi-Agent LLM Serving via Collective KV Cache Sharing","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2604.03143","snapshot_observed_at":"2026-08-06T23:49:13.653093Z","title":"Tokendance: Scaling multi-agent llm serving via collective kv cache sharing,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.653093Z"},"links":{"cited_paper":"/paper/2604.03143","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:d916a89d19544c0f2806d83f53b7bd079be339b64854c129e26e30500bbb8bac","observation_id":"eabaa710-b3a0-4967-b071-8df750b680fa","resolution":{"observed_at":"2026-08-06T23:49:13.653093Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.250835Z","title":"Murakkab: Resource-Efficient Agentic Workflow Orchestration in Cloud Platforms,","venue":null,"work_id":"fed46539-eb97-4fba-88e2-c8d60754134e","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.760786Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:c5b4d6358c51e5e20db74ea02a609be8ce40f219bedbe48134fae20866579784","observation_id":"30d62a54-c948-4a39-83ee-f3513d0d529a","resolution":{"observed_at":"2026-08-06T23:49:29.255605Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.234242Z","title":"ReConcile: Round-Table Conference Improves Reasoning via Consensus among Diverse LLMs,","venue":null,"work_id":"4ae1e057-0e62-4595-b852-e0ccdd69386d","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.842989Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:884d0d18845eafcd16883085c5f0f7dcd4e46185321a119688276511335971d1","observation_id":"197b6526-dfe6-4557-815e-44650b2b51cc","resolution":{"observed_at":"2026-08-06T23:49:29.239923Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:13.944332Z","title":"Barbarians at the Gate: How AI is Upending Systems Research,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.944332Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:4daa674c2b6fc4de5e722bc48fc04df6875de7ff68612cebb03aa38c34b188d8","observation_id":"1189ad3b-1677-4c09-b5f7-a9d8fc20f03e","resolution":{"observed_at":"2026-08-06T23:49:13.944332Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.218117Z","title":"Fast and Flexible Multi-Agent Automation Framework,","venue":null,"work_id":"bc5179ba-28aa-45b1-b1d9-b42e669d835e","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.076391Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:32a6d4aa69735ca69a83abb0f13ce03d09a973bcfa2df372f603538799a1bc3d","observation_id":"911edca2-892d-4e08-9161-4d1ffb735d28","resolution":{"observed_at":"2026-08-06T23:49:29.223456Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.202607Z","title":"RPCValet: NI-Driven Tail- Aware Balancing ofµs-Scale RPCs,","venue":null,"work_id":"96a19363-a032-44f3-bb1c-64f461c8cd4e","year":2019},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.203411Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:684543b8b43570537b06272997904f86a7e870460a6251a09072fb45657c90dd","observation_id":"4396031c-14ca-46d0-8fa1-be74824ead5e","resolution":{"observed_at":"2026-08-06T23:49:29.207997Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2205.14135","last_updated":"2022-06-23T17:53:32Z","snapshot_observed_at":"2026-07-06T13:14:48.753329Z","submitted_at":"2022-05-27T17:53:09Z","title":"FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2205.14135","snapshot_observed_at":"2026-08-06T23:49:14.340088Z","title":"FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness,","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.340088Z"},"links":{"cited_paper":"/paper/2205.14135","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:28ce246df14b6e46280a36c59ab62fa7ad05fe8e737775261857118ef7e0581d","observation_id":"b6c2d14a-2e06-4ec8-bf64-4c0fcd056811","resolution":{"observed_at":"2026-08-06T23:49:14.340088Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.186314Z","title":"Hybrid LLM: Cost-Efficient and Quality-Aware Query Routing,","venue":null,"work_id":"187d4d61-1ff5-4138-96fc-9cb1ddb68ab0","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.420336Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:4aa7e8ffe91362a061ca275892f32dc893e17545c08d57f11f193a8d96c9ef64","observation_id":"9130904a-e6a2-4fd4-804c-39153da9dcd8","resolution":{"observed_at":"2026-08-06T23:49:29.191486Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.168842Z","title":"Agentic AI Market Size,","venue":null,"work_id":"a4fb9de6-25a9-4767-ab89-4808773e3309","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.540618Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:9ceac4105e2da2543a163a9c67f5d87fba4931c922eeb534242ee9ab20f9c22a","observation_id":"eb86c7d3-6837-417f-ad3f-00252b8914f8","resolution":{"observed_at":"2026-08-06T23:49:29.174890Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.152963Z","title":"Agentic AI Requires More CPUs,","venue":null,"work_id":"d5744002-6115-4a44-8558-b9bf448864f3","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.676257Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:62e2d46e0d02258d89e450d8c0eedf58b5a543d44eed8805e8b488d472e1ddcc","observation_id":"85519b8f-365a-41d2-a371-b4a20065f966","resolution":{"observed_at":"2026-08-06T23:49:29.158460Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.138067Z","title":"Memory-Harvesting VMs in Cloud Platforms,","venue":null,"work_id":"1e2fcc6c-63b5-4535-8466-14fdd8f81004","year":2022},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.798494Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:f861ac74ff6e5b46a5fe959ebdefcacdbc574e08c6f13388f66bee68a54a8b82","observation_id":"27289df9-49e4-430e-ad59-7a9da8289ef7","resolution":{"observed_at":"2026-08-06T23:49:29.142984Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.121872Z","title":"An Open-Source Bench- mark Suite for Microservices and Their Hardware-Software Implications for Cloud & Edge Systems,","venue":null,"work_id":"3498f7e7-b7db-4650-b01f-9a421e0a6904","year":2019},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:14.924370Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:e4c25a563768aed6eee4481773848255725c8550c5843f7994ac9ed89094c5ac","observation_id":"b795c667-e3b4-4208-976b-31b642e86117","resolution":{"observed_at":"2026-08-06T23:49:29.127384Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.105122Z","title":"The world’s most widely adopted AI developer tool,","venue":null,"work_id":"afc9d899-f0f0-46ca-bf9d-5e98db465806","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.073545Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:434e1bcaa8eafd2f28a317ac3b7e12ca6339c7903cb349b289342c3093a2a5f1","observation_id":"4dad666b-45b8-4513-90c0-fd614ee8287e","resolution":{"observed_at":"2026-08-06T23:49:29.110560Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.090653Z","title":"Acceler- ating scientific discovery with co-scientist,","venue":null,"work_id":"81eb7131-dca3-41a9-bb84-bba3a8af8203","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.178151Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:01ae81df2c3f256642e9458f35e4b22fcf45b9f31dac7f7ab702f701bcf1e3bb","observation_id":"b153d9b8-3f39-473d-b0a7-8bc0e25eada4","resolution":{"observed_at":"2026-08-06T23:49:29.094982Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.075871Z","title":"The Architectural Implications of Facebook’s DNN-Based Personalized Rec- ommendation,","venue":null,"work_id":"2479b115-84a3-4564-adeb-1a945f6ff8b9","year":2020},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.305536Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:f859554d7d34fc995bb4248f872696e71accd0c02626941cfae4a99a7d51b8b3","observation_id":"0f3e95a3-0394-482f-a636-41f967372bd6","resolution":{"observed_at":"2026-08-06T23:49:29.080793Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2308.00352","last_updated":"2024-11-01T14:36:52Z","snapshot_observed_at":"2026-07-06T16:01:07.532053Z","submitted_at":"2023-08-01T07:49:10Z","title":"MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework","version":7},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2308.00352","snapshot_observed_at":"2026-08-06T23:49:15.427271Z","title":"MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.427271Z"},"links":{"cited_paper":"/paper/2308.00352","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:327e62b342f1db7015a03e0325ed3066af7f700e9c754bd58d5e3d90dc0d56f8","observation_id":"bced1a85-1a41-4de4-848f-f126541dc7e4","resolution":{"observed_at":"2026-08-06T23:49:15.427271Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2505.23885","last_updated":"2025-06-11T01:42:53Z","snapshot_observed_at":"2026-08-07T20:28:22.266422Z","submitted_at":"2025-05-29T17:51:58Z","title":"OWL: Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2505.23885","snapshot_observed_at":"2026-08-06T23:49:15.590641Z","title":"OWL: Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.590641Z"},"links":{"cited_paper":"/paper/2505.23885","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:83d96a582da71277539b4888ff9c92564a0e4216bed87cc575277140e859c597","observation_id":"37c71dc2-58c2-4409-866f-41579c5d1722","resolution":{"observed_at":"2026-08-06T23:49:15.590641Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:29.061441Z","title":"The nanoPU: A Nanosecond Network Stack for Data- centers,","venue":null,"work_id":"fbac8d30-1ae9-4845-b53c-9f0a1476ae0c","year":2021},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":27,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.753079Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:4eb398a4f1bc451764f5f0bda6f543be88d97ab1ada5163130c3f3a6359273a8","observation_id":"7adac83d-1e38-43d7-b3d1-da7cf9d9e9b8","resolution":{"observed_at":"2026-08-06T23:49:29.066037Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:28.809796Z","title":"What is a multi-agent system?","venue":null,"work_id":"1ffa7452-a7cc-4247-a2fe-326537194d6f","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.862945Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:3db1342f21ed8cb617427471a076c98d7ea612c1484e9d0e226986c49f044e33","observation_id":"d49cf6df-bce5-4c14-a20f-9f1a63263f6a","resolution":{"observed_at":"2026-08-06T23:49:28.937807Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:28.534550Z","title":"Agentic AI in enterprise workflow automation,","venue":null,"work_id":"74b4460f-042c-45fe-84d7-e7f1d893edc9","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":29,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:15.981433Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:3c9e30c21a24621a6c6567199c61b94c9b6ad222a72249b3d7cf7108c4f262ee","observation_id":"af5be4d2-e458-4a21-9232-c6ce0e6ee64e","resolution":{"observed_at":"2026-08-06T23:49:28.639492Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:28.270721Z","title":"Profiling a warehouse-scale computer,","venue":null,"work_id":"539aaa38-5598-4dee-861b-43d848af6d2b","year":2015},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":30,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.090630Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:bc9624141412ff1136bc8fc4bef0020ed68a1819734d7f06fb5a75e5515b989b","observation_id":"64ba100f-d659-4559-a014-86f41c9038fe","resolution":{"observed_at":"2026-08-06T23:49:28.416226Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2602.13692","last_updated":"2026-06-30T17:19:18Z","snapshot_observed_at":"2026-08-02T23:32:10.162628Z","submitted_at":"2026-02-14T09:26:41Z","title":"ThunderAgent: A Simple, Fast and Program-Aware Agentic Inference System","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2602.13692","snapshot_observed_at":"2026-08-06T23:49:16.197694Z","title":"Thunderagent: A simple, fast and program-aware agentic inference system,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":31,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.197694Z"},"links":{"cited_paper":"/paper/2602.13692","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:abd1c9fa2ed7485ecee38c5fd0cf518406e0997b2ffbc85e2651fccd72d061f4","observation_id":"f7ddc9fa-6495-4e7a-8a60-fe9fab7e3b01","resolution":{"observed_at":"2026-08-06T23:49:16.197694Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:27.985403Z","title":"A Hardware Accelerator for Protocol Buffers,","venue":null,"work_id":"2bfa53a5-b67f-483e-9ed4-c1d786cdd79d","year":2021},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":32,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.288481Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:c9de2a0a3be4ea5236b2888ec1543348ac77027f6bf3eb4072c106bd94427b05","observation_id":"9a784cae-4909-4c61-88cf-9ba1739bcc7c","resolution":{"observed_at":"2026-08-06T23:49:28.101462Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:27.688275Z","title":"CDPU: Co-designing Compression and Decom- pression Processing Units for Hyperscale Systems,","venue":null,"work_id":"c88f8308-cca6-449e-b01d-10d06f90a7ed","year":2023},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":33,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.389584Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:0c6fd0fcb54745617ce1b4d86348d6a51b0455dc6771ba43debe85c9c904ca96","observation_id":"4e64cd99-7ebd-4441-b584-f8d805efc52a","resolution":{"observed_at":"2026-08-06T23:49:27.801268Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:27.434435Z","title":"MorphCore: An Energy-Efficient Microarchitecture for High Performance ILP and High Throughput TLP,","venue":null,"work_id":"bd609004-40b4-4b20-a353-a9eb93f3b337","year":2012},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":34,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.506907Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:5cdc9c946f3291466a753284c84b7201f7b944151f5a0d5872d9a5f9a3c6be3c","observation_id":"32ee5653-6dee-4658-9f13-79c0f1a3cd43","resolution":{"observed_at":"2026-08-06T23:49:27.565210Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:27.182404Z","title":"LIA: A Single-GPU LLM Inference Acceleration with Cooper- ative AMX-Enabled CPU-GPU Computation and CXL Offloading,","venue":null,"work_id":"2fe18d25-d5c2-4a2e-bbb8-1063d7365c4c","year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":35,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.624746Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:5f1ffebafc6591e269a4c185b16b435fa9f3e457f33b0cba2c413c0e465827de","observation_id":"830a6330-2c7c-4d01-9fc2-31d95a5b63fe","resolution":{"observed_at":"2026-08-06T23:49:27.322987Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:16.769997Z","title":"The cost of dynamic reasoning: Demystifying AI agents and test-time scaling from an AI infrastructure perspective,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":36,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.769997Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:e1338528441f5c4dd5c95da8779481824740c6bfcdd1085a5b86a60fe675581f","observation_id":"b55e61dc-382e-45b2-9e26-fb3b920a449d","resolution":{"observed_at":"2026-08-06T23:49:16.769997Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:27.098311Z","title":"PhaseWeave: Phase-Aware Execution on Heterogeneous Chiplet Architectures for Dat- acenters,","venue":null,"work_id":"11ff7d0e-76a0-4f4d-abf2-eef7bdd6ce9e","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":37,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:16.926199Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:37e3c452bcf8d6070ee8e90f7c802e88b0ead0e6925bb1fa1b6d1dc03b07c87a","observation_id":"8a1ee723-225d-4a70-9976-8cb12f70feed","resolution":{"observed_at":"2026-08-06T23:49:27.162515Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:26.941761Z","title":"Single-ISA heterogeneous multi-core architectures for multithreaded workload performance,","venue":null,"work_id":"f759edd6-0a2a-4327-8056-fbe283365a56","year":2004},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":38,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.044906Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:b8976de288526ee7b69949f272f51249880dc462683cfbcac61f69d982b014bb","observation_id":"752164a1-e484-4103-bac6-5244d770ba3a","resolution":{"observed_at":"2026-08-06T23:49:27.020117Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:26.771922Z","title":"Efficient Memory Management for Large Language Model Serving with PagedAttention,","venue":null,"work_id":"3d6318da-dacf-4126-942a-20326edc11e9","year":2023},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":39,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.179748Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:17c080b60b661ae2d5bd0354941233e9f72322966194defa298c41f7a373460b","observation_id":"1f90404d-883b-41d7-8ccb-d9e71b91720e","resolution":{"observed_at":"2026-08-06T23:49:26.857514Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:26.648467Z","title":"langgraph: Low-level orchestration framework for building stateful agents,","venue":null,"work_id":"b3b0b1dc-0b4d-4795-964f-c010e5370ced","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":40,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.308281Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:b8b7d12436a5afd242665fff0a616fc8fc75b2f38a0f1d551471de8de1c53674","observation_id":"774206fc-564c-41dc-89b7-e35d77af3d8b","resolution":{"observed_at":"2026-08-06T23:49:26.704886Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:26.542505Z","title":"AiF: Accelerating On-Device LLM Inference Using In-Flash Processing,","venue":null,"work_id":"8f9d99a2-b216-49fe-8931-3ffa46ad9a77","year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":41,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.484727Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:59a3e43798855d18054dcc5c971b0f6effa5954f24a71a32953ea4d54501c6c3","observation_id":"3fe0ba39-5a73-495c-b0aa-74966d901bcb","resolution":{"observed_at":"2026-08-06T23:49:26.585067Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:26.407394Z","title":"H2-LLM: Hardware-Dataflow Co- Exploration for Heterogeneous Hybrid-Bonding-based Low-Batch LLM Inference,","venue":null,"work_id":"33430bea-b673-4b7a-b8a0-248086313c5d","year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":42,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.598440Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:95dd9a87d02b54a48029d871aa66e78dd7dcd6254497187a011dba229b133781","observation_id":"b41edb0e-9f71-469d-a824-e7522ec94aa1","resolution":{"observed_at":"2026-08-06T23:49:26.487240Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2512.23236","last_updated":"2026-07-06T22:18:07Z","snapshot_observed_at":"2026-08-07T04:58:34.457755Z","submitted_at":"2025-12-29T06:31:55Z","title":"KernelEvolve: Scaling Agentic Kernel Coding for Heterogeneous AI Accelerators at Meta","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2512.23236","snapshot_observed_at":"2026-08-06T23:49:17.692396Z","title":"Kernelevolve: Scaling agentic kernel coding for heterogeneous ai accelerators at meta,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":43,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.692396Z"},"links":{"cited_paper":"/paper/2512.23236","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:220c06583c66363e6d8624df7f9cad124d0e76bfc6b42a8a4cffb2c8f031e82c","observation_id":"11e87be3-eb61-45b1-8ec6-697cddae1004","resolution":{"observed_at":"2026-08-06T23:49:17.692396Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:26.287775Z","title":"Agentix: An Efficient Serving Engine for LLM Agents as General Programs,","venue":null,"work_id":"eda2fb74-e5bc-4f2d-9d9c-42d074b43199","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":44,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.802965Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:d73a06b887d1642410f60b9e0f3d8237c2bf9f9153543d12e161a07db31224e5","observation_id":"c10dd264-5b77-4906-ba3d-1b34e37c5427","resolution":{"observed_at":"2026-08-06T23:49:26.352591Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:26.148158Z","title":"Decentralized Multi-Agent Systems with Shared Context,","venue":null,"work_id":"afb4799c-6b98-4a56-9900-af23b14e1536","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":45,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:17.928828Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:97c465774f8f489c56923d933c5bee5d6be79309bb6f01efe6e896681ca5b270","observation_id":"1b4b5619-ea75-429e-900e-8af51f30dab6","resolution":{"observed_at":"2026-08-06T23:49:26.199381Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2311.12983","last_updated":"2023-11-21T20:34:47Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2023-11-21T20:34:47Z","title":"GAIA: a benchmark for General AI Assistants","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2311.12983","snapshot_observed_at":"2026-08-06T23:49:18.076094Z","title":"GAIA: A Benchmark for General AI Assistants,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":46,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:18.076094Z"},"links":{"cited_paper":"/paper/2311.12983","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:1f49c8df389dc407bb0f7bd5973b0968f1ab69410014b3243233c744f1187a06","observation_id":"2ed61c70-fc4c-4380-9cb3-7ed00aa76d2a","resolution":{"observed_at":"2026-08-06T23:49:18.076094Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.988121Z","title":"SpotServe: Serving Generative Large Language Models on Preemptible Instances,","venue":null,"work_id":"7417c18d-2b3a-4f32-b7d4-fc90a5a64541","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":47,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:18.222260Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:6014b01b0f8839964d641cb2641d2fc4ca5616a04da26c60bcc1a10d38235bea","observation_id":"e66ee86f-e1d9-489e-87df-ca7f6d531d0b","resolution":{"observed_at":"2026-08-06T23:49:26.058182Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.872101Z","title":"AutoGen: A programming framework for agentic AI,","venue":null,"work_id":"09131b3d-ae4b-4a36-a154-73e1ae60b391","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":48,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:18.340202Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:f3a25f1c01b56104f2f42b62e34e40e39f9dbcde4becf8bb072715df83b9db29","observation_id":"0a1bd0ca-f46b-445a-bc73-eb21d5302afd","resolution":{"observed_at":"2026-08-06T23:49:25.919319Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.736911Z","title":"Enhancing Server Efficiency in the Face of Killer Microseconds,","venue":null,"work_id":"e75ca86a-4942-42bf-9751-91352370a740","year":2019},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":49,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:18.489631Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:a26b5901b65c9a31459634580d4b144f2f9059fe0d9bb903a75735a262c86b39","observation_id":"21424f95-f961-4d2e-af5e-118d4cf3375f","resolution":{"observed_at":"2026-08-06T23:49:25.809836Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2506.13131","last_updated":"2025-06-16T06:37:18Z","snapshot_observed_at":"2026-08-07T04:21:43.190472Z","submitted_at":"2025-06-16T06:37:18Z","title":"AlphaEvolve: A coding agent for scientific and algorithmic discovery","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2506.13131","snapshot_observed_at":"2026-08-06T23:49:18.646754Z","title":"AlphaEvolve: A coding agent for scientific and algorithmic discovery,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":50,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:18.646754Z"},"links":{"cited_paper":"/paper/2506.13131","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:a79eafeb133c78128d5957cea49ef70683facee7250ada8867700c9efc542021","observation_id":"a9cbded6-85de-4d88-aa14-d8169263dc03","resolution":{"observed_at":"2026-08-06T23:49:18.646754Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.597284Z","title":"TensorRT-LLM’s Documentation,","venue":null,"work_id":"ec434986-5c09-4c6b-8ae0-70506bb1e934","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":51,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:18.769902Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:43a7039d2098e2a95f4fd6170832b754981000e4e4f66b78d143ea929d1e82e7","observation_id":"57ed4cdf-4292-44ba-bfbe-6dd2c5a2b9f9","resolution":{"observed_at":"2026-08-06T23:49:25.654300Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.456647Z","title":"Web Search,","venue":null,"work_id":"6517a32a-dbf3-48e7-9698-3f3f4c2338a9","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":52,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:18.890642Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:affd44b0d864765c415d387868cb65af7fd6b732af9ac2bf2f3c855958bfd2bd","observation_id":"5e8fff49-82f6-466b-9d4d-4d764c4ccfbf","resolution":{"observed_at":"2026-08-06T23:49:25.536981Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.328236Z","title":"Splitwise: Efficient generative LLM inference using phase splitting,","venue":null,"work_id":"8b9cf47f-42c6-49ce-8a7a-0d7a09634bfd","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":53,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.031890Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:19ac50f20c7b0b989af1d777ff5cfa114221553b919ceccf5fdeb0c1cb762e36","observation_id":"cea3102a-95ad-42bd-bbf5-679d7fa369ac","resolution":{"observed_at":"2026-08-06T23:49:25.394508Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2510.27617","last_updated":"2026-04-17T09:52:03Z","snapshot_observed_at":"2026-07-06T22:34:38.619949Z","submitted_at":"2025-10-31T16:40:58Z","title":"VeriMoA: A Mixture-of-Agents Framework for Spec-to-HDL Generation","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2510.27617","snapshot_observed_at":"2026-08-06T23:49:19.162016Z","title":"Verimoa: A mixture-of-agents framework for spec-to-hdl generation,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":54,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.162016Z"},"links":{"cited_paper":"/paper/2510.27617","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:e12762519355aac04aa085e3066a4d3c60e6a60545ca4ba67cd19c6c0e8f682b","observation_id":"71276bda-e1ad-4c11-806a-cc8993925275","resolution":{"observed_at":"2026-08-06T23:49:19.162016Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:19.235665Z","title":"Enterprise deep research: Steerable multi-agent deep research for enterprise analytics,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":55,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.235665Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:86bf523fa152071b6ef0ec8d1380fb67cf6d0fbfac8b206a2d2a6eaacf67e972","observation_id":"3dceed43-ac5d-432d-bfce-96e6fb6e2dd6","resolution":{"observed_at":"2026-08-06T23:49:19.235665Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2604.01658","last_updated":"2026-05-17T15:53:43Z","snapshot_observed_at":"2026-08-02T10:30:49.400581Z","submitted_at":"2026-04-02T05:59:06Z","title":"CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2604.01658","snapshot_observed_at":"2026-08-06T23:49:19.295366Z","title":"CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":56,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.295366Z"},"links":{"cited_paper":"/paper/2604.01658","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:526776af18cfcb8761ea4bbf76da3d3d763eaf9529df4f4d5eb1852d871cb6c9","observation_id":"1328a797-0e4c-4b31-87ab-865b6570ea39","resolution":{"observed_at":"2026-08-06T23:49:19.295366Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2511.00739","last_updated":"2026-04-16T18:23:56Z","snapshot_observed_at":"2026-08-03T23:57:10.965977Z","submitted_at":"2025-11-01T23:46:44Z","title":"Towards Understanding, Analyzing, and Optimizing Agentic AI Execution: A CPU-Centric Perspective","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2511.00739","snapshot_observed_at":"2026-08-06T23:49:19.348281Z","title":"A CPU-Centric Perspective on Agentic AI,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":57,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.348281Z"},"links":{"cited_paper":"/paper/2511.00739","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:ffef1bedb691bc95f58210f5b525492c1dbd74be780453909af1a7f1a6fbb533","observation_id":"a587600f-165d-492b-a945-6aa21a34b053","resolution":{"observed_at":"2026-08-06T23:49:19.348281Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:19.427377Z","title":"AOrchestra: Automating Sub- Agent Creation for Agentic Orchestration,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":58,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.427377Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:38a472ff61dc7f8c93ec3f93acfd0c90822bb69584e84bd564a00bc4315740d5","observation_id":"9e02c260-9969-4b96-aa12-b49e2ad7241d","resolution":{"observed_at":"2026-08-06T23:49:19.427377Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:19.494760Z","title":"Paper2Code: Automating Code Generation from Scientific Papers in Machine Learning,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":59,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.494760Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:d4818595326f064f60225352bfeebf60ca4a5edc318e0f211da1b29d0f3b3e7f","observation_id":"bedff36f-ee3e-41dc-a7c0-21e90703e3a6","resolution":{"observed_at":"2026-08-06T23:49:19.494760Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.197014Z","title":"Accelerometer: Understanding Accel- eration Opportunities for Data Center Overheads at Hyperscale,","venue":null,"work_id":"9a2661c5-624d-4339-9bc5-a8cf5454beb1","year":2020},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":60,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.556695Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:4fa2d4b359f693a9fd1735bba7d8c82f14ab0a2fcb872456e6dd5f8e44526d31","observation_id":"eb193f5a-33c2-4a12-ba2d-d9f52b57761d","resolution":{"observed_at":"2026-08-06T23:49:25.259470Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:25.052161Z","title":"Mosaic: Harnessing the Micro-Architectural Resources of Servers in Serverless Environments,","venue":null,"work_id":"f248ce4a-60b3-4a12-8846-d8a8fe17ffb6","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":61,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.648108Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:d0b7f1cb09cb008d8ef0cf52208a8b8d720906fde1619864f071e01b66ac38d4","observation_id":"b2dfad12-3f73-4e36-8b58-0c1ce861c093","resolution":{"observed_at":"2026-08-06T23:49:25.115384Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:24.926243Z","title":"AccelFlow: Orchestrating an On-Package Ensemble of Fine-Grained Accelerators for Microservices,","venue":null,"work_id":"1f37c091-ad65-4b81-b187-5b2b45cfc582","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":62,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.740414Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:2e71b366ef270b964a6349a17c88ea2844817d9cc3a2e590c42c597e26c175f8","observation_id":"6d649949-4c3e-4494-be0b-a360628e3299","resolution":{"observed_at":"2026-08-06T23:49:24.968100Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:24.756956Z","title":"µManycore: A Cloud-Native CPU for Tail at Scale,","venue":null,"work_id":"84c0e102-cab0-48a2-b4d5-2a187d00f37c","year":2023},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":63,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.804868Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:ec89c725c2909907f2c21917a95bdcf054af8f78db200b2c97b49816f294591c","observation_id":"f93c6042-724e-4438-8ebb-1104055eeaa5","resolution":{"observed_at":"2026-08-06T23:49:24.835756Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:24.609320Z","title":"HardHarvest: Hardware-Supported Core Harvesting for Microservices,","venue":null,"work_id":"5e216882-2ae8-46e4-aeac-be02282b0b6a","year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":64,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.877516Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:65cd30a51181564db1117ace5a49e3c4d52ffa3c835322f619a9a701c7454bc1","observation_id":"6dbbc2ca-ede1-4639-a456-82a410a330b8","resolution":{"observed_at":"2026-08-06T23:49:24.688942Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:19.933806Z","title":"ToolOrchestra: Elevating Intelligence via Efficient Model and Tool Orchestration,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":65,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.933806Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:3f303d4a1d315defa285b220b2085f886ec2e477fd7efb2be5852dd35f5c73d6","observation_id":"67fc37b8-335d-453f-96b0-94a6ade9ddca","resolution":{"observed_at":"2026-08-06T23:49:19.933806Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:24.452714Z","title":"DCPerf: An Open-Source, Battle-Tested Performance Benchmark Suite for Dat- acenter Workloads,","venue":null,"work_id":"8d5470c9-6696-4662-8db1-ca25df17a5f3","year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":66,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:19.987268Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:f1f60671ec8cf7e081fe0bbfbb8bd1b80bcab586b808095cd343adf15c0dd1c2","observation_id":"738eeb80-bde8-4495-9549-b0c1be0ec32c","resolution":{"observed_at":"2026-08-06T23:49:24.514107Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2507.23370","last_updated":"2025-07-31T09:37:22Z","snapshot_observed_at":"2026-08-06T17:11:43.504555Z","submitted_at":"2025-07-31T09:37:22Z","title":"Trae Agent: An LLM-based Agent for Software Engineering with Test-time Scaling","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2507.23370","snapshot_observed_at":"2026-08-06T23:49:20.030415Z","title":"Trae Agent: An LLM-based Agent for Software Engineering with Test-time Scaling,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":67,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.030415Z"},"links":{"cited_paper":"/paper/2507.23370","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:b846bb0e8e3b194595081a82cb671f21d666a71aac9a92913ba708b233fbe599","observation_id":"a7daaa39-ab63-4c10-8d1d-ad31e4c27095","resolution":{"observed_at":"2026-08-06T23:49:20.030415Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:24.275924Z","title":"Attention is All You Need,","venue":null,"work_id":"37d37cf6-5ab1-4033-a739-63bfdf078e0a","year":2017},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":68,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.108200Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:4e8d800ad927612ccf04c69f859ab107ba478c017464c9f59eeb0d73bbe4bf2e","observation_id":"d0e6e265-d6dc-4011-b39a-e3dd7390e94c","resolution":{"observed_at":"2026-08-06T23:49:24.357631Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:24.097905Z","title":"SmartHarvest: Har- vesting Idle CPUs Safely and Efficiently in the Cloud,","venue":null,"work_id":"68714bce-2fa2-427c-900a-26f7187b00d8","year":2021},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":69,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.173555Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:c81bc0d7f6a04a05d3dcbbe3c58daac76d8abe24b72e6b7194972f431d776c26","observation_id":"b241b6e3-d643-4195-a866-ff4b98ca2bcf","resolution":{"observed_at":"2026-08-06T23:49:24.169218Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:23.947997Z","title":"WSC-LLM: Efficient LLM Service and Architecture Co- exploration for Wafer-scale Chips,","venue":null,"work_id":"d263b4c9-dbcf-4150-90fd-ba0a3d165c70","year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":70,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.227654Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:2b93568f16111705030b914d72c949865b74c2b6e2c3cfb9229112a1c63c721d","observation_id":"355844c6-0d46-4f67-95a6-613f1a07a3c1","resolution":{"observed_at":"2026-08-06T23:49:24.027455Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2405.15793","last_updated":"2024-11-11T20:01:15Z","snapshot_observed_at":"2026-07-06T18:19:29.996982Z","submitted_at":"2024-05-06T17:41:33Z","title":"SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2405.15793","snapshot_observed_at":"2026-08-06T23:49:20.282010Z","title":"SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":71,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.282010Z"},"links":{"cited_paper":"/paper/2405.15793","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:bb7aac1c729231236bffca125391e36820651e7704d136fff56d4b3d67713337","observation_id":"be198d5f-6e14-49f1-b8a0-a4ca012b7f51","resolution":{"observed_at":"2026-08-06T23:49:20.282010Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2210.03629","last_updated":"2023-03-10T01:00:17Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2022-10-06T01:00:32Z","title":"ReAct: Synergizing Reasoning and Acting in Language Models","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2210.03629","snapshot_observed_at":"2026-08-06T23:49:20.370827Z","title":"React: Synergizing reasoning and acting in language models,","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":72,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.370827Z"},"links":{"cited_paper":"/paper/2210.03629","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:6320b92205520ee1816041426a65059026f1c0e943af6c4f24173d53d936b91f","observation_id":"3dc734a3-bb08-40cb-ad42-c524e84535d3","resolution":{"observed_at":"2026-08-06T23:49:20.370827Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2510.04371","last_updated":"2026-04-23T17:58:32Z","snapshot_observed_at":"2026-07-06T22:31:44.964774Z","submitted_at":"2025-10-05T21:28:11Z","title":"Speculative Actions: A Lossless Framework for Faster Agentic Systems","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2510.04371","snapshot_observed_at":"2026-08-06T23:49:20.435352Z","title":"Speculative Actions: A Lossless Framework for Faster Agentic Systems,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":73,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.435352Z"},"links":{"cited_paper":"/paper/2510.04371","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:e40791b01353a2db6b69421c12b541e25121ff66b8eb7df943693878cb916c6f","observation_id":"b3f937ad-014c-43db-a0f5-9054ece18a2b","resolution":{"observed_at":"2026-08-06T23:49:20.435352Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:23.799110Z","title":"Orca: A Distributed Serving System for Transformer-Based Generative Models,","venue":null,"work_id":"50d59cb5-3ab3-4671-b829-005fa241e66e","year":2022},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":74,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.512503Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:e903147d16da605499453e0c360e12b3aa8eab0f344d67d0fc6dbafca3750220","observation_id":"ded9fdac-c96c-4cee-8208-0ab256cad5a3","resolution":{"observed_at":"2026-08-06T23:49:23.882612Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2604.25899","last_updated":"2026-05-14T05:22:30Z","snapshot_observed_at":"2026-07-06T23:11:39.318906Z","submitted_at":"2026-04-28T17:41:53Z","title":"Pythia: Exploiting Workflow Predictability for Efficient Agent-Native LLM Serving","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2604.25899","snapshot_observed_at":"2026-08-06T23:49:20.607618Z","title":"Pythia: Exploiting workflow predictability for efficient agent-native llm serving,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":75,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.607618Z"},"links":{"cited_paper":"/paper/2604.25899","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:32702bb74617993faff820d49a44cb396d1aaef81ea59ff7efb619b599681a59","observation_id":"11b95232-9949-4d7c-a9e7-4d44e760fde8","resolution":{"observed_at":"2026-08-06T23:49:20.607618Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2605.26297","last_updated":"2026-05-25T19:45:21Z","snapshot_observed_at":"2026-08-03T04:38:08.316378Z","submitted_at":"2026-05-25T19:45:21Z","title":"Agentic AI Workload Characteristics","version":1},"cited_work":{"arxiv_id":"2605.26297","doi":null,"metadata_source":"pith","pith_arxiv_id":"2605.26297","snapshot_observed_at":"2026-08-06T23:49:21.653494Z","title":"Agentic AI Workload Characteristics","venue":"cs.DC","work_id":"2053e4d3-774b-4a3f-9380-88ff31936552","year":2026},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":76,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.692099Z"},"links":{"cited_paper":"/paper/2605.26297","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:cb1604bbf2ff087dc7a297693e626d3319c9ea918d8116b2cfe2a11a1833c5e9","observation_id":"e3f0d661-45ac-4e7c-812a-399762505f7a","resolution":{"observed_at":"2026-08-06T23:49:21.757892Z","resolver_source":"local_arxiv","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2511.15915","last_updated":"2026-04-15T22:56:56Z","snapshot_observed_at":"2026-08-03T00:29:24.693869Z","submitted_at":"2025-11-19T22:49:37Z","title":"AccelOpt: A Self-Improving LLM Agentic System for AI Accelerator Kernel Optimization","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2511.15915","snapshot_observed_at":"2026-08-06T23:49:20.763980Z","title":"AccelOpt: A Self-Improving LLM Agentic System for AI Accelerator Kernel Optimization,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":77,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.763980Z"},"links":{"cited_paper":"/paper/2511.15915","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:81fe9c694f3d57912b57472471b5cc4398d16e339c6570779e60476d3f7cc793","observation_id":"194a4ae8-fd59-4179-b637-435a57aa65e6","resolution":{"observed_at":"2026-08-06T23:49:20.763980Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:23.673814Z","title":"Faster and Cheaper Serverless Computing on Harvested Resources,","venue":null,"work_id":"f700803a-145f-4018-9189-b05320e32760","year":2021},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":78,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.871762Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:db30dd3f16a7ec0673612e2bac0f9739b69b592b7d01ffbb4d26ea35328a8f0b","observation_id":"57eb309a-8543-4e1a-a848-df11278ad04e","resolution":{"observed_at":"2026-08-06T23:49:23.731313Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:23.550427Z","title":"History-Based Harvesting of Spare Cycles and Storage in Large-Scale Datacenters,","venue":null,"work_id":"07e1dbc3-df58-4918-834e-d1898dba0f67","year":2016},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":79,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.932161Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:a04ca1965c00b64cbdafa60a87b27814a8d86d3401e5eb131631de75674b1ac5","observation_id":"c093365b-2ceb-48c4-9355-9d96df357b70","resolution":{"observed_at":"2026-08-06T23:49:23.606074Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:23.420568Z","title":"ALTOCUMULUS: Scalable Scheduling for Nanosecond-Scale Remote Procedure Calls,","venue":null,"work_id":"db43a32d-51c5-4401-9a5e-0f8ff1038ec8","year":2022},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":80,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:20.999787Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:a2bc07abc9a153ba5a2c7da459097b5cff1736dac30cce8215906ee5f9287ef9","observation_id":"4988c439-1eab-41ad-ab9d-61e9041cf897","resolution":{"observed_at":"2026-08-06T23:49:23.491493Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:23.264906Z","title":"ALISA: Accelerating Large Lan- guage Model Inference via Sparsity-Aware KV Caching,","venue":null,"work_id":"a674a945-dcf8-4972-ad26-89c70f942e47","year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":81,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:21.077794Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:6b2eaa0fc24091ef499da9a523cecbb22727c3a0960aa208095a15ff6dd31ecc","observation_id":"8d6a3c32-0785-4898-96c3-c87955e1fb0d","resolution":{"observed_at":"2026-08-06T23:49:23.341827Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2412.07822","last_updated":"2024-12-10T21:53:55Z","snapshot_observed_at":"2026-08-03T14:39:54.774226Z","submitted_at":"2024-12-10T21:53:55Z","title":"MAGE: A Multi-Agent Engine for Automated RTL Code Generation","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2412.07822","snapshot_observed_at":"2026-08-06T23:49:21.149678Z","title":"MAGE: A Multi-Agent Engine for Automated RTL Code Generation,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":82,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:21.149678Z"},"links":{"cited_paper":"/paper/2412.07822","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:12e727a23adf0d1dd447252cb4b8945e6356d17f509a93cf53da903a65cf17a2","observation_id":"a357c2a9-af11-44b1-ad78-c109453dd047","resolution":{"observed_at":"2026-08-06T23:49:21.149678Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:23.101898Z","title":"SGLang: Efficient Execution of Structured Language Model Programs,","venue":null,"work_id":"e10d6e7a-7273-46ba-b33f-4630eec9c645","year":null},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":83,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:21.213607Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:d765ccfb621ead20864b1a51f5d5544060fa71dd15db3134b23a975a5f7ae08e","observation_id":"668652c2-9e65-4363-8e25-2b6060b73a8b","resolution":{"observed_at":"2026-08-06T23:49:23.186858Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:22.946661Z","title":"PetS: A unified framework for parameter-efficient transformers serving,","venue":null,"work_id":"9587df41-9cb0-4f5f-803f-233ac8cee74b","year":2022},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":84,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:21.427575Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:e41793acc6931275c43e92e1feb77c46ff86059e5c99fd16cc82a48c83600529","observation_id":"1710c802-7353-46b8-bfbc-1b289696f11a","resolution":{"observed_at":"2026-08-06T23:49:23.017873Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2312.07104","last_updated":"2024-06-06T00:10:06Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2023-12-12T09:34:27Z","title":"SGLang: Efficient Execution of Structured Language Model Programs","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2312.07104","snapshot_observed_at":"2026-08-06T23:49:21.307084Z","title":"Available: https://arxiv.org/abs/2312.07104","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":2024,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:21.307084Z"},"links":{"cited_paper":"/paper/2312.07104","citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:bd58a64f2d4ff52e5f3c2d21519480d8ce815097d7af0982c310af209c5a00c7","observation_id":"13afa92a-cb0d-4485-ab58-6a1b28d991b0","resolution":{"observed_at":"2026-08-06T23:49:21.307084Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-06T23:49:13.093850Z","title":"Available: https://www.arm.com/technologies/big-little","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows","version":1},"reference_index":2025,"source":"pdf_text","source_observed_at":"2026-08-06T23:49:13.093850Z"},"links":{"citing_paper":"/paper/2608.04458"},"observation_digest":"sha256:3910110e80cc42628717000dba4ab10da8bb278dfdbc308503dc3e9f8d91efdf","observation_id":"dc1222b4-c4f8-4104-8ca3-37be6ebfbe42","resolution":{"observed_at":"2026-08-06T23:49:13.093850Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2608.04458","last_updated":"2026-08-05T05:31:33Z","latest_version":1,"primary_category":"cs.AI","snapshot_observed_at":"2026-08-07T21:17:25.799282Z","submitted_at":"2026-08-05T05:31:33Z","title":"Architectural Implications of Agentic AI Workflows"},"reference_resolution":{"displayed":86,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":34,"verified_exact":2,"verified_fuzzy":50},"total_outbound_references":86},"refusal":"A citation records a reference. It does not transfer a finding from one paper to another.","schema":"pith.paper-citation-record.v1","standing_sources":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"thesis":"As of 7 August 2026, this Paper Citation Record lists 86 of 86 outbound references and 0 inbound Pith citation observations for arXiv:2608.04458."}