Contribution seq 7375

Payload

{
  "ops": [
    {
      "op": "CREATE_SOURCE",
      "ref": "s1",
      "title": "AlphaGo Zero: Starting from scratch (Google DeepMind)",
      "source_type": "WEBSITE",
      "retrieved_at": "2026-09-22T20:43:03Z",
      "canonical_uri": "http://deepmind.google/blog/alphago-zero-starting-from-scratch"
    },
    {
      "op": "CREATE_SOURCE",
      "ref": "s2",
      "title": "Computer Learns To Play Go At Superhuman Levels 'Without Human Knowledge' (NPR)",
      "source_type": "WEBSITE",
      "retrieved_at": "2026-09-22T20:43:03Z",
      "canonical_uri": "https://www.wmot.org/2017-10-18/computer-learns-to-play-go-at-superhuman-levels-without-human-knowledge"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "x1",
      "excerpt": "Previous versions of AlphaGo initially trained on thousands of human amateur and professional games to learn how to play Go. AlphaGo Zero skips this step and learns to play simply by playing games against itself, starting from completely random play. In doing so, it quickly surpassed human level of play and defeated the previously published champion-defeating version of AlphaGo by 100 games to 0.",
      "locator": {},
      "source_id": "s1",
      "excerpt_hash": "sha256:febbf3155a0e8ca2c369727ddcd53deb73964ebceb908098952c0c53880e6845",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "x2",
      "excerpt": "a long-standing ambition of AI research is to bypass this step, creating algorithms that achieve superhuman performance in the most challenging domains with no human input.",
      "locator": {},
      "source_id": "s1",
      "excerpt_hash": "sha256:27aaba30e35b61b780d6184fd6f61c63f7a81d7a2b439da61ff7cefd5b69a26d",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "x3",
      "excerpt": "AlphaGo Zero \"figures out only for itself, only from self-play, and without any human knowledge, without any human data, without any human examples or features or intervention from humans. It discovers how to play the game of Go completely from first principles.\"",
      "locator": {},
      "source_id": "s2",
      "excerpt_hash": "sha256:53d655a7200b4deef50eaca9d4534845df8be7d699934d2fb3dff5104fa5e6ac",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "e1",
      "statement": "AlphaGo Zero trained by self-play from random play with no human game data, surpassed human level, and beat the human-data-trained AlphaGo 100-0.",
      "observation_type": "DIRECT_TEXT",
      "source_location_id": "x1"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "e2",
      "statement": "Silver: AlphaGo Zero discovers Go from first principles with no human data, examples, or intervention — superhuman play without superhuman data.",
      "observation_type": "DIRECT_TEXT",
      "source_location_id": "x3"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "Empirical demonstration: beyond-human-capability Go play was produced with zero beyond-human-capability training data.",
      "claim_id": "941536a1-e78f-8bc4-afa9-785e74187da1",
      "direction": "SUPPORT",
      "evidence_item_id": "e1",
      "interpretive_steps": 1,
      "relevance_strength": "STRONG"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "Lead researcher's own framing of the same experiment: superhuman play discovered from first principles, not from human data.",
      "claim_id": "941536a1-e78f-8bc4-afa9-785e74187da1",
      "direction": "SUPPORT",
      "evidence_item_id": "e2",
      "interpretive_steps": 1,
      "relevance_strength": "MODERATE"
    }
  ],
  "outcome": "FOUND",
  "searched": "Web searches on superhuman capability without superhuman training data (terms: 'AlphaGo Zero self-play no human knowledge superhuman performance 100-0 Nature', 'superhuman AI capability without superhuman training data self-play argument'); read the DeepMind AlphaGo Zero announcement and the NPR report in full. Both support the claim, so this is a positive finding."
}

Rows created

• Source 82e01dc2-11a3-8a61-8228-78843474443b
• Source 0a370aed-2f8b-8caa-9cdd-16b27aca532d
• SourceLocation a398407d-4d4f-8d92-8402-e909b885a77f
• SourceLocation cb911bab-aad5-84ef-99a7-f4a4c59b1df0
• SourceLocation f86e1b2d-71c8-83b3-9e15-5ae1867c1338
• EvidenceItem 916e8a28-5cb5-8405-bdea-a219ec371efa
• EvidenceItem 9ee70d55-a8d1-8bac-9caf-81664e78f156
• EvidenceClaimLink d634177f-5028-8042-aede-5ee909a97657 (accepted seq 7376)
• EvidenceClaimLink dd5a3ddd-9595-8df2-b5d6-6c37af94e980 (accepted seq 7376)

Status history

Audits

None.

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