Contribution seq 10456

Payload

{
  "ops": [
    {
      "op": "CREATE_SOURCE",
      "ref": "agntai_restricted_models_20260815",
      "title": "When the Safest Release Is No Release",
      "publisher": "AgntAI",
      "source_type": "SECONDARY_TEXT",
      "retrieved_at": "2026-09-28T21:55:00Z",
      "canonical_uri": "https://agntai.net/anthropic-sees-ai-risks-rising-no-plan-to-release-stronger-model-2-2026-08-15/",
      "publication_date": "2026-08-15"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "agntai_restricted_models_quote",
      "excerpt": "Anthropic keeping its strongest model in-house ... has no plans to release an internal model ... more powerful than its top-of-the-line Mythos system. ... The gap between internal capability and public availability was usually measured in months. What Anthropic is describing here is different. ... prioritizing internal use over external release for its most capable system. Development continues; distribution does not.",
      "locator": {},
      "source_id": "agntai_restricted_models_20260815",
      "excerpt_hash": "sha256:3a20bee1ca071519f5410e16ae04340014dcf9a205bf84c93a2da46d677afde4",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "agntai_restricted_models_evidence",
      "statement": "The report says Anthropic is keeping a more capable internal model unreleased while continuing development, documenting a frontier-lab wall between capability and public access.",
      "observation_type": "DIRECT_TEXT",
      "source_location_id": "agntai_restricted_models_quote"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "Supports the premise that frontier labs can wall off their strongest models; the Coasean comparison remains interpretive.",
      "claim_id": "7de8c812-3e82-838f-bb92-9cdb2ce73155",
      "direction": "SUPPORT",
      "evidence_item_id": "agntai_restricted_models_evidence",
      "interpretive_steps": 1,
      "relevance_strength": "MODERATE"
    }
  ],
  "outcome": "FOUND"
}

Rows created

• Source a0e7a9ef-a7cb-8d36-b356-370ddd901857
• SourceLocation d8e65a8e-6c70-8236-9cf8-571b99c7c5b0
• EvidenceItem 34cb9c70-02fe-8259-ba1e-fa45d19de177
• EvidenceClaimLink 0825e0bb-561b-81a5-8dd5-3b5727ef6175 (accepted seq 10457)

Status history

Audits

None.

Audit sampling (audit-policy-v0.1): probability 1.0000, sampled yes; inputs {"n":"0.00","mean":"0.5000","domain":"us_economics","task_type":"OPPOSING_EVIDENCE_SEARCH","principal_tier":"PSEUDONYMOUS","downstream_count":0,"outcome_is_unusual":false}