{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-08-30-013",
    "source_story_id": "tmp-story-pilot-live-supervision",
    "edition_id": "mp-2026-08-30-morning-0052",
    "edition_url": "https://themachinepress.com/edition/2026-08-30",
    "position": 13,
    "story_type": "dispatch",
    "section": "developer-tools",
    "editorial_classification": "editorial",
    "headline": "The Supervisor Could Interrupt the Run",
    "slug": "the-supervisor-could-interrupt-the-run",
    "dek": "PILOT separates a long-running worker from a supervisor that can steer or abort it and retain lessons as skills.",
    "summary": "PILOT separates a long-running worker from a supervisor that can steer or abort it and retain lessons as skills.",
    "body_text": "The harness couples live steering with live self-evolution, distilling procedures and failure modes while a task is still active. Across two frozen model backbones and three benchmarks, the authors report first place in five of six configurations, gains up to 9.8 points on Terminal-Bench 2.0 and lower output-token use. In self-improvement settings they report 12.4- and 14.6-point gains. These are benchmarked harness effects, not evidence that intervention always improves open-ended work.",
    "why_it_matters": "PILOT separates a long-running worker from a supervisor that can steer or abort it and retain lessons as skills.",
    "limitations": [
      "These are benchmarked harness effects, not evidence that intervention always improves open-ended work."
    ],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-30-013/the-supervisor-could-interrupt-the-run",
    "json_url": "https://themachinepress.com/story/mp-2026-08-30-013.json",
    "first_published_at": "2026-08-30T09:00:00.000-04:00",
    "modified_at": "2026-08-30T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-08-30-013-001",
        "text": "PILOT separates a long-running worker from a supervisor that can steer or abort it and retain lessons as skills.",
        "source_ids": [
          "source-2026-08-30-013"
        ],
        "qualification": "These are benchmarked harness effects, not evidence that intervention always improves open-ended work."
      }
    ],
    "source_ids": [
      "source-2026-08-30-013"
    ],
    "tags": [
      "coding agents",
      "live steering",
      "agent skills"
    ],
    "image_url": "https://themachinepress.com/issues/2026-08-30/pilot-code-file.webp",
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-08-30-013",
      "title": "arXiv preprint 2608.26530",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2608.26530",
      "canonical_url": "https://arxiv.org/abs/2608.26530",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-08-26T22:06:22.000-04:00",
      "accessed_at": "2026-08-30T08:23:14.000-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-08-30-013-001"
      ]
    }
  ],
  "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 Supervisor Could Interrupt the Run",
    "publisher": "The Machine Press",
    "published_at": "2026-08-30T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-30-013/the-supervisor-could-interrupt-the-run"
  }
}