{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreicq6jw5oawwamreltovqiabeygp3c3vbchyqqz7udfiodekzom7iq",
    "uri": "at://did:plc:avkh7zze5iapdkk6naaunrjn/app.bsky.feed.post/3ml7agoibjh72"
  },
  "path": "/260506/p38#a260506p38",
  "publishedAt": "2026-05-06T17:00:54.000Z",
  "site": "https://www.techmeme.com",
  "tags": [
    "Ars Technica",
    "Google releases Multi-Token Prediction drafters for its Gemma 4 models, which use a form of speculative decoding to guess future tokens for faster inference"
  ],
  "textContent": "Ryan Whitwam / Ars Technica:\n**Google releases Multi-Token Prediction drafters for its Gemma 4 models, which use a form of speculative decoding to guess future tokens for faster inference** — Google launched its Gemma 4 open models this spring, promising a new level of power and performance for local AI.",
  "title": "Google releases Multi-Token Prediction drafters for its Gemma 4 models, which use a form of speculative decoding to guess future tokens for faster inference (Ryan Whitwam/Ars Technica)"
}