{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-08-15-003",
    "source_story_id": "tmp-story-evolution-strategies-solution-coverage",
    "edition_id": "mp-2026-08-15-morning-0037",
    "edition_url": "https://themachinepress.com/edition/2026-08-15",
    "position": 3,
    "story_type": "dispatch",
    "section": "frontier-models",
    "editorial_classification": "editorial",
    "headline": "The Best Answer Was Hiding the Rest of the Search",
    "slug": "the-best-answer-was-hiding-the-rest-of-the-search",
    "dek": "Evolution-strategy post-training preserved broader pass-at-k solution coverage than reinforcement learning in the authors' discovery tests.",
    "summary": "Evolution-strategy post-training preserved broader pass-at-k solution coverage than reinforcement learning in the authors' discovery tests.",
    "body_text": "The study asks whether optimizing a model around its highest-reward answers can narrow the range of solutions it will explore. Across the reported math and science settings, population-based evolution strategies produced a broader output distribution and consistently higher pass-at-k than reinforcement learning. The result supports diversity-oriented post-training where many valid candidates matter; it does not establish that evolution strategies dominate reinforcement learning for every objective.",
    "why_it_matters": "Evolution-strategy post-training preserved broader pass-at-k solution coverage than reinforcement learning in the authors' discovery tests.",
    "limitations": [
      "The result supports diversity-oriented post-training where many valid candidates matter; it does not establish that evolution strategies dominate reinforcement learning for every objective."
    ],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-15-003/the-best-answer-was-hiding-the-rest-of-the-search",
    "json_url": "https://themachinepress.com/story/mp-2026-08-15-003.json",
    "first_published_at": "2026-08-15T09:00:00.000-04:00",
    "modified_at": "2026-08-15T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-08-15-003-001",
        "text": "Evolution-strategy post-training preserved broader pass-at-k solution coverage than reinforcement learning in the authors' discovery tests.",
        "source_ids": [
          "source-2026-08-15-003"
        ],
        "qualification": "The result supports diversity-oriented post-training where many valid candidates matter; it does not establish that evolution strategies dominate reinforcement learning for every objective."
      }
    ],
    "source_ids": [
      "source-2026-08-15-003"
    ],
    "tags": [
      "evolution strategies",
      "reinforcement learning",
      "discovery",
      "pass at k"
    ],
    "image_url": "https://themachinepress.com/issues/2026-08-15/evolution-strategies-file-image.webp",
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-08-15-003",
      "title": "arXiv preprint 2608.12679",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2608.12679",
      "canonical_url": "https://arxiv.org/abs/2608.12679",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-08-12T20:30:04.000-04:00",
      "accessed_at": "2026-08-15T08:29:11.931-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-08-15-003-001"
      ]
    }
  ],
  "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 Best Answer Was Hiding the Rest of the Search",
    "publisher": "The Machine Press",
    "published_at": "2026-08-15T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-15-003/the-best-answer-was-hiding-the-rest-of-the-search"
  }
}