{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-09-04-026",
    "source_story_id": "tmp-story-gated-deltanet-full-fp4",
    "edition_id": "mp-2026-09-04-morning-0057",
    "edition_url": "https://themachinepress.com/edition/2026-09-04",
    "position": 15,
    "story_type": "dispatch",
    "section": "chips-infrastructure",
    "editorial_classification": "editorial",
    "headline": "The Recurrent Half Was the Easy Half to Quantize",
    "slug": "the-recurrent-half-was-the-easy-half-to-quantize",
    "dek": "Minima applies NVFP4 weights and activations to all 496 linear layers of a hybrid 27B model, gates included.",
    "summary": "Minima applies NVFP4 weights and activations to all 496 linear layers of a hybrid 27B model, gates included.",
    "body_text": "Community recipes had left Gated DeltaNet layers—especially decay and write-strength gates—at higher precision because recurrent error was expected to accumulate. The tested 17.5-GiB recipe instead matched BF16 within seed noise across its suite, improved prefill by 14% to 19% and kept retrieval through 64K. Mechanism studies found block scaling localized outliers, gate nonlinearities compressed error and the delta rule forgot injected disturbances. The checkpoint and claims apply to the evaluated Qwen3.8-27B hybrid.",
    "why_it_matters": "Minima applies NVFP4 weights and activations to all 496 linear layers of a hybrid 27B model, gates included.",
    "limitations": [],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-09-04-026/the-recurrent-half-was-the-easy-half-to-quantize",
    "json_url": "https://themachinepress.com/story/mp-2026-09-04-026.json",
    "first_published_at": "2026-09-04T09:00:00.000-04:00",
    "modified_at": "2026-09-04T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-09-04-026-001",
        "text": "Minima applies NVFP4 weights and activations to all 496 linear layers of a hybrid 27B model, gates included.",
        "source_ids": [
          "source-2026-09-04-015"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-09-04-015"
    ],
    "tags": [
      "quantization",
      "Gated DeltaNet",
      "LLM inference"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-09-04-015",
      "title": "arXiv preprint 2609.04098",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2609.04098",
      "canonical_url": "https://arxiv.org/abs/2609.04098",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-09-03T13:04:26.000-04:00",
      "accessed_at": "2026-09-04T08:24:03.294-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-09-04-026-001"
      ]
    }
  ],
  "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 Recurrent Half Was the Easy Half to Quantize",
    "publisher": "The Machine Press",
    "published_at": "2026-09-04T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-09-04-026/the-recurrent-half-was-the-easy-half-to-quantize"
  }
}