{
  "title": "The V7 Research Strands",
  "story": [
    {
      "type": "markdown",
      "id": "e8af7358c045d95c",
      "text": "# 🔭 The V7 Research Strands"
    },
    {
      "type": "markdown",
      "id": "0aabeac16e89a316",
      "text": "The model's seventh pass, at **headline level**. Each strand is a reading in progress: named here, kept in the atlas, and published in full only when its note is ready. Nothing on this page extends a proven result; conjecture status lives in the register."
    },
    {
      "type": "reference",
      "id": "bf4128d2f57b40f1",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "v7-public-agreement-and-private-ambiguity",
      "title": "Observer-patch holography",
      "text": "An outside physics framework read against the model: where public agreement and private ambiguity line up with the Swordsman/Mage split, and where the analogy must not be carried."
    },
    {
      "type": "reference",
      "id": "7b9b76b6d9d4b27c",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "v7-merkle-tree-certificates-and-the-hold",
      "title": "Merkle Tree Certificates × the Hold",
      "text": "Batch-signed certificates as a reading of what the Star holds: one signed root, many inclusion proofs."
    },
    {
      "type": "reference",
      "id": "c36ae9cb2927e871",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "v7-unlinkable-inference-and-the-swordsman",
      "title": "Unlinkable inference",
      "text": "Private inference as a Swordsman concern: requests that cannot be linked back to the asker."
    },
    {
      "type": "reference",
      "id": "bbbd55cccaae3245",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "v7-what-the-verifier-reads-siggolf",
      "title": "What the verifier reads",
      "text": "A hash-signature board as a lens on disclosure: what a verifier must read, and what it never needs."
    },
    {
      "type": "reference",
      "id": "28c003ef46f1107d",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "v7-openai-math-research-letters",
      "title": "Research letters on the reconstruction ceiling",
      "text": "Read against outside mathematics, the strict R < 1 is **declared, not derived**: the capacity deficit is a premise a deployment declares and attests."
    },
    {
      "type": "reference",
      "id": "a701717db096afd3",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "the-declared-deficit-r-1-declared-not-derived",
      "title": "The declared deficit",
      "text": "The ruling itself, as it now stands in the dragon skill and the model page."
    },
    {
      "type": "reference",
      "id": "b4220255dfe4cfcd",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "v7-embedding-reconstruction-and-measured-disclosure",
      "title": "Embeddings and measured disclosure",
      "text": "How much of a text an embedding gives back, treated as measured disclosure rather than assumed privacy."
    },
    {
      "type": "reference",
      "id": "6a530dacb4836088",
      "site": "atlas.guide.agentprivacy.ai",
      "slug": "v7-evidence-authority-and-effects",
      "title": "Evidence, authority and effects",
      "text": "A loop for agent actions: what evidence an action carries, who authorised it, and what effect it may have."
    },
    {
      "type": "markdown",
      "id": "a54e9e9f2886ba2c",
      "text": "## Navigation\n← [[Welcome Visitors]] · [[The Research Notes]]"
    }
  ],
  "journal": [
    {
      "type": "create",
      "item": {
        "title": "The V7 Research Strands",
        "story": [
          {
            "type": "markdown",
            "id": "e8af7358c045d95c",
            "text": "# 🔭 The V7 Research Strands"
          },
          {
            "type": "markdown",
            "id": "0aabeac16e89a316",
            "text": "The model's seventh pass, at **headline level**. Each strand is a reading in progress: named here, kept in the atlas, and published in full only when its note is ready. Nothing on this page extends a proven result; conjecture status lives in the register."
          },
          {
            "type": "reference",
            "id": "bf4128d2f57b40f1",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "v7-public-agreement-and-private-ambiguity",
            "title": "Observer-patch holography",
            "text": "An outside physics framework read against the model: where public agreement and private ambiguity line up with the Swordsman/Mage split, and where the analogy must not be carried."
          },
          {
            "type": "reference",
            "id": "7b9b76b6d9d4b27c",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "v7-merkle-tree-certificates-and-the-hold",
            "title": "Merkle Tree Certificates × the Hold",
            "text": "Batch-signed certificates as a reading of what the Star holds: one signed root, many inclusion proofs."
          },
          {
            "type": "reference",
            "id": "c36ae9cb2927e871",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "v7-unlinkable-inference-and-the-swordsman",
            "title": "Unlinkable inference",
            "text": "Private inference as a Swordsman concern: requests that cannot be linked back to the asker."
          },
          {
            "type": "reference",
            "id": "bbbd55cccaae3245",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "v7-what-the-verifier-reads-siggolf",
            "title": "What the verifier reads",
            "text": "A hash-signature board as a lens on disclosure: what a verifier must read, and what it never needs."
          },
          {
            "type": "reference",
            "id": "28c003ef46f1107d",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "v7-openai-math-research-letters",
            "title": "Research letters on the reconstruction ceiling",
            "text": "Read against outside mathematics, the strict R < 1 is **declared, not derived**: the capacity deficit is a premise a deployment declares and attests."
          },
          {
            "type": "reference",
            "id": "a701717db096afd3",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "the-declared-deficit-r-1-declared-not-derived",
            "title": "The declared deficit",
            "text": "The ruling itself, as it now stands in the dragon skill and the model page."
          },
          {
            "type": "reference",
            "id": "b4220255dfe4cfcd",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "v7-embedding-reconstruction-and-measured-disclosure",
            "title": "Embeddings and measured disclosure",
            "text": "How much of a text an embedding gives back, treated as measured disclosure rather than assumed privacy."
          },
          {
            "type": "reference",
            "id": "6a530dacb4836088",
            "site": "atlas.guide.agentprivacy.ai",
            "slug": "v7-evidence-authority-and-effects",
            "title": "Evidence, authority and effects",
            "text": "A loop for agent actions: what evidence an action carries, who authorised it, and what effect it may have."
          },
          {
            "type": "markdown",
            "id": "a54e9e9f2886ba2c",
            "text": "## Navigation\n← [[Welcome Visitors]] · [[The Research Notes]]"
          }
        ]
      },
      "date": 1791331200000
    }
  ]
}