{"_canonicalization":{"envelope_id":"axm_ + sha256(envelope minus {signature, axiom_id, anchors})","envelope_signature":"ed25519(envelope minus {signature, axiom_id})","json":"sort_keys=True, separators=(',',':'), ensure_ascii=False, allow_nan=False, utf-8","leaf_hash":"sha256(0x00 || canonical_json(envelope_full))","seal_signature":"ed25519(seal minus {signature, sig_algorithm})"},"axiom_id":"axm_b7d9b2215bb065d24767cc7b5ae14906930b3d841fcc98b0a79315171723d452","bitcoin_anchor":{"bitcoin_attestations":[],"calendar_attestations":[],"ots_url":"","stamped_at":"","status":"pending_next_stamp"},"envelope":{"anchors":[{"chain":"crovia.axiom_graph","height":0,"merkle_proof":"spider_vendor_press_v1","root_at_anchor":"spider_vendor_press_v1"}],"axiom_id":"axm_b7d9b2215bb065d24767cc7b5ae14906930b3d841fcc98b0a79315171723d452","axiom_type":"AX.OBS","body":{"axiom_subtype":"news.vendor_press.v1","category":"news","fingerprint":"14659ca51a79d6cd4df8d56d7d380078dc8c62a2679cfb1193f06649328dfc30","published":"Tue, 21 Jul 2026 00:00:00 -0400","receipt_hash":"14659ca51a79d6cd4df8d56d7d380078dc8c62a2679cfb1193f06649328dfc30","schema":"spider.news.vendor_press.v1","spider":"vendor_press","spider_record":{"axiom_subtype":"news.vendor_press.v1","category":"news","decision_hint":"POSITIVE","envelope_target":"AX.OBS","fingerprint":"14659ca51a79d6cd4df8d56d7d380078dc8c62a2679cfb1193f06649328dfc30","observed_at":"2026-07-21T04:43:35.036805Z","parent_run_hash":"03e944014de2697434479833d15ea9303e014945afc230ecc7f207824493b589","published":"Tue, 21 Jul 2026 00:00:00 -0400","runtime_version":"0.1.0","schema":"spider.news.vendor_press.v1","source_status":200,"source_url":"https://export.arxiv.org/rss/cs.AI","spider":"vendor_press","summary_excerpt":"arXiv:2607.17136v1 Announce Type: cross \nAbstract: Agentic computer-use RL is reported in single runs, and those numbers mislead. Using verifier-guided repair of a 35B computer-use agent (CUA) across five oracle-graded environments, we show a repaired policy's success rate is dominated by upstream variance: a variance-components decomposition across three cells (crossed data-draw $\\times$ seed grid, bootstrap CIs) finds evaluation variance negligible ($\\sigma_{\\mathrm{eval}} \\approx 0$) and the training-seed effect small everywhere ($\\leq 10\\%$); instead it splits between the data draw and run-to-run nondeterminism, the data draw's share rising to dominant ($48\\%$) on the hardest cell. There the run-to-run distribution is bimodal (Hartigan dip $p=0.07$, $k=10$), so a single run has roughly a 30% chance of the failure mode and mean$\\pm$std is the wrong summary. On that footing, two findings hold. First, repairability is two-tier in how constrained the corrective action is: a single fixe","title":"Teach it to stop, not just to click","url":"https://arxiv.org/abs/2607.17136","vendor":"arxiv_cs_ai"},"summary":"arXiv:2607.17136v1 Announce Type: cross \nAbstract: Agentic computer-use RL is reported in single runs, and those numbers mislead. Using verifier-guided repair of a 35B computer-use agent (CUA) across five oracle-graded environments, we show a repaired policy's success rate is dominated by upstream variance: a variance-components decomposition across three cells (crossed data-draw $\\times$ seed grid, bootstrap CIs) finds evaluation variance negligible ($\\sigma_{\\mathrm{eval}} \\approx 0$) and the training-seed effect small everywhere ($\\leq 10\\%$); instead it splits between the data draw and run-to-run nondeterminism, the data draw's share rising to dominant ($48\\%$) on the hardest cell. There the run-to-run distribution is bimodal (Hartigan dip $p=0.07$, $k=10$), so a single run has roughly a 30% chance of the failure mode and mean$\\pm$std is the wrong summary. On that footing, two findings hold. First, repairability is two-tier in how constrained the corrective action is: a single fixe","title":"Teach it to stop, not just to click","vendor":"arxiv_cs_ai"},"confidence":{"method":"deterministic"},"decision":"POSITIVE","issued_at":"2026-07-21T04:43:35Z","notes":"Spider vendor_press (news) news.vendor_press.v1","object":{"captured_by":"crovia.spider.vendor_press","primary_source_url":"https://arxiv.org/abs/2607.17136"},"predecessors":[],"schema":"crovia.axiom.v1","signature":"ed25519:ee36eeacee64e36eec6793043512f7e37c49d9feafd25ef92875dce188e3e269b7edf521f5c0e733cdc16210e52316132bca0de682bc3579bde7004e4fa38c0b","signer":"crovia.substrate","subject":{"observed_at":"2026-07-21T04:43:35Z","source_collector":"spider:vendor_press","target_id":"https://arxiv.org/abs/2607.17136"},"tsa":{"authority":"crovia.substrate.bootstrap","rfc3161_token":"{\"kind\":\"crovia.bootstrap.tsa\",\"source_jsonl\":\"/opt/crovia/spider/data/news/vendor_press_v1.jsonl\",\"source_seal_merkle_root\":\"spider_vendor_press_v1\",\"upgrade_path\":\"Sessione H \\u2014 OpenTimestamps weekly anchor\"}"},"zk_mode":"clear","zk_proof":null},"ledger":{"leaf_hash":"983b0e729a259cc0b023f29cb2a4b2d86eaf4a2c2849f8a8c779d8f01f986faa","leaf_index":336747,"ledger_path":"/opt/crovia/substrate/axiom_ledger.jsonl"},"merkle_proof":{"hash_alg":"sha256","leaf_prefix":"0x00","node_prefix":"0x01","odd_leaf_rule":"duplicate_last","path":[{"sibling":"2799bf503ba23fb9ec8fb5f3c843752137d6edd7af34cf155af5c8c69ec73d8c","side":"left"},{"sibling":"f65b1d1372f7ba4e12d343f14d068df70ae1d8c60dcb3a32f99bd6453f0684b6","side":"left"},{"sibling":"c1e3bc662a47decb037e8a353048cb77fdb164af0aa6fe0b4a5e903bb2509c54","side":"right"},{"sibling":"40df37a07ba93606fb0e86b600ebb677924d4dfd7d2faadb9724bd06d031604f","side":"left"},{"sibling":"4f6645b27feed6996386bc3a6f0c774107613beb8de492a0533cfd2387d2d65d","side":"right"},{"sibling":"0f11a14b52dbd7ce194a8ec5a9532d618ad4ab1219f25bbd53b7eae1c8d17760","side":"left"},{"sibling":"c3453c4355287f8f6ea5b33a2f2d95f482fa4ede383b4407cc8d044e332ee8e5","side":"left"},{"sibling":"a4c41a614872a391d764732d35b746845ccfe3f2a51fe13e1f0e934d271e94f4","side":"right"},{"sibling":"bd233cee8c0876447824d86de33608c36e3a8e17dca0737155291ed768cf56d1","side":"left"},{"sibling":"7b927551b5db06b6571913b4e6792ffcce5291a3eca5a0df4a3b6e296271105f","side":"left"},{"sibling":"b77a0b5ae4607c8fe6ba73449d46b35076e3dedc0c82a2c65a05780d42a7bc2e","side":"right"},{"sibling":"9eb5077edfb3dc553857d4794b925bfce117e0f8a1d049af5d0dd9026b470eef","side":"right"},{"sibling":"414b1a70fd1dcb25489a194714b97492b066684b15d0b7a48a176c4b9b5bc713","side":"right"},{"sibling":"21d66dd41003813f710b7617944f1bfba3258658a5d3370c21cad8f9e945bc99","side":"left"},{"sibling":"613f015699131eb89bd755dee67133be95af25cf5f16c1c8ce4b99d963b8dd86","side":"right"},{"sibling":"a729b574b1135956436ded5eef1fe8f08014ff6a0729749d307ab1bca93fcdc9","side":"right"},{"sibling":"77025bcb374a7ad74f520643e20a8ae1205a7ee78507b0beb93117760f1c29d3","side":"left"},{"sibling":"4bf21052e085e8ac81f1dec1d2b310bd12bf948992de6177d12e9d2fda8d39f0","side":"right"},{"sibling":"1cecb7f447febd025aac272837c80de218aecc6485d2395a509b2a1f1b9c746e","side":"left"}]},"schema":"crovia.axiom_proof.v1","seal":{"first_collector_run_id":"","first_receipt_hash":"","jsonl_path":"/opt/crovia/substrate/axiom_ledger.jsonl","key_id":"430895f101d38164","last_collector_run_id":"","last_receipt_hash":"","leaf_count":337144,"merkle_root":"5e969cc01afa67e4dbe5d37b712cdb10f4aa1fd74404e02eab724cf487c8d6d9","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","run_id":"hourly_json_retrofit_20260721T053701Z","schema":"crovia.seal.v1","seal_family_version":"crovia-seal-family/1","seal_kind":"substrate_batch","sealed_at":"2026-07-21T05:38:38Z","sig_algorithm":"ed25519","signature":"5c0c1a8dd2793d787ccd5e49e8b4d70eed136352555518589f05c74be357fc171702e42a76c3a556d90e3d51ff36cb3d292aaac83c66566de7b943f318bda50c","signer_version":"1.1.0"},"trust_root":{"key_id":"430895f101d38164","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","signature_algorithm":"ed25519","url":"/registry/canon/TRUST_ROOT.md"},"verifier":{"spec":"/registry/canon/AXIOM_RECEIPT_v1.md","url":"/v/axm_b7d9b2215bb065d24767cc7b5ae14906930b3d841fcc98b0a79315171723d452"}}