{
  "schema_version": "1.1",
  "id": "archive:https://arxiv.org/abs/2611.09848v1",
  "slug": "2611-09848v1-1h2pt4o",
  "url": "https://feed7.dev/p/2611-09848v1-1h2pt4o",
  "title": "How far does a hash-only sender go? A pre-registered map of sender memory versus bits for short text messages",
  "why_included": "A pre-registered test finds the hash-only sender costs more bits than an in-domain PPM sender once failed deliveries need fallback text. The study mainly maps a narrow compression tradeoff, not an agent workflow.",
  "summary": "On **1,000** Taskmaster-1 utterances, the hash-only sender delivered 857 on the first try and cost **192.8 bits per message** for complete delivery, 18.4 bits more than PPM order 1. Recovering half the failures with a wider receiver search lowered the bit cost by 7.7 bits versus PPM, but took about three minutes of receiver time per failure.",
  "practical_implication": "For builders choosing compact text transmission, account for the full retry path: acknowledgements and raw-text fallback can erase the initial savings. The sender's memory-versus-bits curve also matters if the device has a strict memory budget.",
  "agent_context": "On **1,000** Taskmaster-1 utterances, the hash-only sender delivered 857 on the first try and cost **192.8 bits per message** for complete delivery, 18.4 bits more than PPM order 1. Recovering half the failures with a wider receiver search lowered the bit cost by 7.7 bits versus PPM, but took about three minutes of receiver time per failure.\n\nFor builders choosing compact text transmission, account for the full retry path: acknowledgements and raw-text fallback can erase the initial savings. The sender's memory-versus-bits curve also matters if the device has a strict memory budget.\n\nThis is one pre-registered run on English utterances from one dataset. Airtime comparisons vary by link, and the wider receiver search still loses to PPM on four of five priced links.",
  "source": {
    "name": "arXiv",
    "url": "https://arxiv.org/abs/2611.09848v1",
    "published_at": "2026-11-06T18:31:24.000Z"
  },
  "source_class": "blog_post",
  "content_type": "Paper",
  "layer": "industry",
  "domains": [
    "data"
  ],
  "topics": [],
  "verification": {
    "status": "needs_review",
    "label": "Needs Review",
    "method": "unverified",
    "verified_at": null
  },
  "uncertainty": [
    "This is one pre-registered run on English utterances from one dataset. Airtime comparisons vary by link, and the wider receiver search still loses to PPM on four of five priced links."
  ],
  "connected_context": {
    "meaning": "This quantifies how retry and receiver-search costs can erase the initial savings of hash-only text transmission, with results varying by link and a single English dataset. It reinforces evaluating total delivery cost under the actual device and channel constraints rather than judging compression by first-pass bits alone.",
    "corpus_size": 1032,
    "generated_at": "2026-11-09T11:01:38.055Z",
    "connections": [
      {
        "title": "Balancing hierarchical representation and quantization: A performance trade-off in text embeddings",
        "source_name": "arXiv",
        "source_url": "https://arxiv.org/abs/2611.06859v1",
        "feed7_url": "https://feed7.dev/p/2611-06859v1-1tqbhix",
        "reason": "Both measure a compression trade-off against task quality or cost: the embedding study tracks storage and retrieval effectiveness, while this study includes retries and receiver time in transmission cost."
      },
      {
        "title": "Measure Before You Train: Typed Decision Models vs. a Classical Baseline vs. a Generative LLM for Routing a Personal Multilingual Inbox",
        "source_name": "arXiv",
        "source_url": "https://arxiv.org/abs/2610.32955v1",
        "feed7_url": "https://feed7.dev/p/2610-32955v1-18vy54w",
        "reason": "Both show that a headline metric can miss the operational objective: classification accuracy misses actionable-email costs, while first-pass bit savings omit retry and fallback costs."
      }
    ]
  },
  "lifecycle": "Current",
  "published_at": "2026-11-06T18:31:24.000Z",
  "modified_at": "2026-11-06T18:31:24.000Z",
  "supersedes": [],
  "expires_at": null,
  "formats": {
    "html": "https://feed7.dev/p/2611-09848v1-1h2pt4o",
    "json": "https://feed7.dev/p/2611-09848v1-1h2pt4o.json",
    "markdown": "https://feed7.dev/p/2611-09848v1-1h2pt4o.md"
  }
}