{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-08-14-003",
    "source_story_id": "tmp-story-latent-self-distillation",
    "edition_id": "mp-2026-08-14-morning-0036",
    "edition_url": "https://themachinepress.com/edition/2026-08-14",
    "position": 3,
    "story_type": "dispatch",
    "section": "frontier-models",
    "editorial_classification": "editorial",
    "headline": "The Teacher Learned What Privilege to Remember",
    "slug": "the-teacher-learned-what-privilege-to-remember",
    "dek": "Latent on-policy self-distillation made the teacher's private context learnable and used less than thirty percent of two comparison methods' rollout budgets.",
    "summary": "Latent on-policy self-distillation made the teacher's private context learnable and used less than thirty percent of two comparison methods' rollout budgets.",
    "body_text": "LOPD retrieves relevant experiences, compresses them into continuous latent tokens for a privileged self-teacher and supplies dense token-level supervision along the student's own trajectories. On agentic tool-use and code-generation tasks, the authors report gains over reinforcement learning with verifiable rewards and several self-distillation baselines. Ablations attribute the improvement to learning the privileged context instead of prescribing answers, feedback or skills. The evidence is benchmark performance, not a demonstration of autonomous open-ended self-improvement.",
    "why_it_matters": "Latent on-policy self-distillation made the teacher's private context learnable and used less than thirty percent of two comparison methods' rollout budgets.",
    "limitations": [
      "The evidence is benchmark performance, not a demonstration of autonomous open-ended self-improvement."
    ],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-14-003/the-teacher-learned-what-privilege-to-remember",
    "json_url": "https://themachinepress.com/story/mp-2026-08-14-003.json",
    "first_published_at": "2026-08-14T09:00:00.000-04:00",
    "modified_at": "2026-08-14T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-08-14-003-001",
        "text": "Latent on-policy self-distillation made the teacher's private context learnable and used less than thirty percent of two comparison methods' rollout budgets.",
        "source_ids": [
          "source-2026-08-14-003"
        ],
        "qualification": "The evidence is benchmark performance, not a demonstration of autonomous open-ended self-improvement."
      }
    ],
    "source_ids": [
      "source-2026-08-14-003"
    ],
    "tags": [
      "self-distillation",
      "agents",
      "tool use",
      "training efficiency"
    ],
    "image_url": "https://themachinepress.com/issues/2026-08-14/latent-self-distillation-file-image.webp",
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-08-14-003",
      "title": "arXiv preprint 2608.13040",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2608.13040",
      "canonical_url": "https://arxiv.org/abs/2608.13040",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-08-13T06:05:51.000-04:00",
      "accessed_at": "2026-08-14T08:27:00.000-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-08-14-003-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "The Teacher Learned What Privilege to Remember",
    "publisher": "The Machine Press",
    "published_at": "2026-08-14T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-14-003/the-teacher-learned-what-privilege-to-remember"
  }
}