Step 2: tool results

get_claim

{
  "include": [
    "arguments",
    "subclaims",
    "dependents"
  ],
  "claim_id": "72ac75a9-6f75-45b9-997c-205f081655f1"
}

{"claim":{"id":"72ac75a9-6f75-45b9-997c-205f081655f1","kind":"claim","canonical_form":"AI agents outperform human experts on short AI research-engineering tasks but underperform them on longer tasks.","claim_type":"empirical_verifiable","state":"active","decomposition_status":"complete","importance":0.35,"contestation":0.3,"domains":[],"tags":[{"slug":"ai-research-automation","name":"AI Research Automation"},{"slug":"ai-agents","name":"AI Agents"},{"slug":"ai-capability-evaluation","name":"AI Capability Evaluation"}],"steward_state":"running","stewarded_at":"2026-11-13T21:45:24.311Z","canonical_direction_note":null,"created_at":"2026-11-05T00:07:17.088Z","updated_at":"2026-11-13T19:55:31.290Z","page_url":"https://minerval.ai/claims/72ac75a9-6f75-45b9-997c-205f081655f1"},"assessment":null,"coverage":{"examined":false,"on":[],"note":"No judgment of this node is in force. That is not a finding either way."},"preliminary_seed":{"credence":0.8,"note":"METR's RE-Bench (Wijk et al., 2024; ICML 2025) found frontier agents beat human experts at a 2-hour budget but scored below them at 8 hours and beyond. The pattern was measured on seven environments with late-2024 models, so whether it persists for newer models is the open part.","seeded_by_claim_id":"ceb833ca-a085-4b45-aba3-8bd3d58add9c"},"quantity":null,"threshold_of":null,"question":null,"formalization":null,"formalization_pending":null,"formalization_history":[],"verification":null,"lean_checks":[],"lean_checks_total":0,"waiting_messages":[{"trigger":"structure_and_assess","from":"the system","sender_node_id":null,"sent_at":"2026-11-09T19:00:56.960Z","paid":false,"summary":"[structure_and_assess] New claim onboarded. Structure it (decompose, calling match_claim for each dependency to link existing claims or create new ones), then assess it."},{"trigger":"mandate_plan","from":"a mandate's plan","sender_node_id":null,"sent_at":"2026-11-13T19:55:31.292Z","paid":true,"summary":"A funded mandate scheduled a assess pass on this claim."}],"requests":[],"requests_total":0,"subclaim_count":0,"instance_count":0,"source_map":null,"subclaims":[],"subclaims_assessed":0,"dependents":[{"id":"ceb833ca-a085-4b45-aba3-8bd3d58add9c","text":"The ability to sustain autonomous work on long tasks is the main capability bottleneck preventing AI from automating AI research.","claim_type":"empirical_derived","node_kind":"claim","relation_type":"supports","reasoning":"If agents already match experts at short durations and fall behind only as the budget grows, the deficit is located in sustaining work over time, which is what the claim names as the bottleneck.","importance":0.45,"assessment_status":"doubtful","assessment_confidence":0.55,"assessment_credence":0.3,"bounty_micro_usd":null,"checked":null,"formal":false}],"dependents_total":1,"arguments":[]}