{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreihyb7cfp77xoaqob3scoxmh7qcr3zstyx4yk7xlibrc7pwvonwk7a",
    "uri": "at://did:plc:2m7awgzbxz4cwloms7hvf6el/app.bsky.feed.post/3mgvgnpmuzpz2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreidhoxfx5pyh4hlgvlcufu3t3vf6oxto7kk7ydj6lsunuq4plxlfau"
    },
    "mimeType": "image/jpeg",
    "size": 203660
  },
  "path": "/google-gemini-embedding-2-launch/",
  "publishedAt": "2026-03-12T11:45:49.000Z",
  "site": "https://gadgetbond.com",
  "tags": [
    "AI",
    "Google",
    "Tech",
    "Gemini AI (formerly Bard)",
    "Source"
  ],
  "textContent": "Google’s new embedding model maps five different media types into one semantic space, so a single query can cut across your entire content stack.\n\nSource",
  "title": "Google’s new Gemini Embedding 2 supercharges multimodal RAG"
}