{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:OHT6LHTPQMD5EICVSVAMEDHFGA","short_pith_number":"pith:OHT6LHTP","schema_version":"1.0","canonical_sha256":"71e7e59e6f8307d220559540c20ce5303a508f1978e94dee1d2887b7f1907e79","source":{"kind":"arxiv","id":"2608.11323","version":1},"attestation_state":"computed","paper":{"title":"Deployment Decision Reliability: A Generalizability-Theory Framework for Sizing Long-Horizon Agent Evaluations","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Vasundra Srinivasan","submitted_at":"2026-08-11T18:16:51Z","abstract_excerpt":"Enterprise practitioners read agent leaderboards as if they ranked agent capability. We show, across three open agent-trace benchmarks (TheAgentCompany, $\\tau^2$-bench, and AppWorld), that the agent main effect accounts for less than 3% of total variance in every dataset and check type, while the agent-by-task interaction accounts for 7-23%. Leaderboards rank specialization, not capability. We arrive at this through a four-facet Generalizability Theory variance decomposition, fit with three estimators (Henderson Method-I, REML via lme4, and a Bayesian binomial GLMM) that agree to three decimal"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2608.11323","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-11T18:16:51Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"7d0cb65c3d291d84155a15fc8de85311f1349199a4d931ea6d6fea178c97c54f","abstract_canon_sha256":"ca8d51413a37ad4d98762ff9e8f07be68dd5c40148aa0c5d9c4306e427f86e55"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-13T00:21:35.571484Z","signature_b64":"4QJoKoeY+54BviaiVrKW2ib0R/gkk+JW/99cv9uUwLtUZdKqs5eUqlK4msXu1DUtBjINuvbxoN8LIV62y1I4AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"71e7e59e6f8307d220559540c20ce5303a508f1978e94dee1d2887b7f1907e79","last_reissued_at":"2026-08-13T00:21:35.569699Z","signature_status":"signed_v1","first_computed_at":"2026-08-13T00:21:35.569699Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deployment Decision Reliability: A Generalizability-Theory Framework for Sizing Long-Horizon Agent Evaluations","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Vasundra Srinivasan","submitted_at":"2026-08-11T18:16:51Z","abstract_excerpt":"Enterprise practitioners read agent leaderboards as if they ranked agent capability. We show, across three open agent-trace benchmarks (TheAgentCompany, $\\tau^2$-bench, and AppWorld), that the agent main effect accounts for less than 3% of total variance in every dataset and check type, while the agent-by-task interaction accounts for 7-23%. Leaderboards rank specialization, not capability. We arrive at this through a four-facet Generalizability Theory variance decomposition, fit with three estimators (Henderson Method-I, REML via lme4, and a Bayesian binomial GLMM) that agree to three decimal"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.11323","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2608.11323/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2608.11323","created_at":"2026-08-13T00:21:35.570183+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.11323v1","created_at":"2026-08-13T00:21:35.570183+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.11323","created_at":"2026-08-13T00:21:35.570183+00:00"},{"alias_kind":"pith_short_12","alias_value":"OHT6LHTPQMD5","created_at":"2026-08-13T00:21:35.570183+00:00"},{"alias_kind":"pith_short_16","alias_value":"OHT6LHTPQMD5EICV","created_at":"2026-08-13T00:21:35.570183+00:00"},{"alias_kind":"pith_short_8","alias_value":"OHT6LHTP","created_at":"2026-08-13T00:21:35.570183+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA","json":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA.json","graph_json":"https://pith.science/api/pith-number/OHT6LHTPQMD5EICVSVAMEDHFGA/graph.json","events_json":"https://pith.science/api/pith-number/OHT6LHTPQMD5EICVSVAMEDHFGA/events.json","paper":"https://pith.science/paper/OHT6LHTP"},"agent_actions":{"view_html":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA","download_json":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA.json","view_paper":"https://pith.science/paper/OHT6LHTP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.11323&json=true","fetch_graph":"https://pith.science/api/pith-number/OHT6LHTPQMD5EICVSVAMEDHFGA/graph.json","fetch_events":"https://pith.science/api/pith-number/OHT6LHTPQMD5EICVSVAMEDHFGA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA/action/storage_attestation","attest_author":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA/action/author_attestation","sign_citation":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA/action/citation_signature","submit_replication":"https://pith.science/pith/OHT6LHTPQMD5EICVSVAMEDHFGA/action/replication_record"}},"created_at":"2026-08-13T00:21:35.570183+00:00","updated_at":"2026-08-13T00:21:35.570183+00:00"}