{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreigf22lg36iosnxdillftuasjeufl7hxexaagytxr3bpwkxbumhxim",
"uri": "at://did:plc:2gbt2dlwaqovtnmxkat3tyke/app.bsky.feed.post/3mmsn4ok3ycj2"
},
"coverImage": {
"$type": "blob",
"ref": {
"$link": "bafkreifrpercoeyogxx2ltt5m5dznc2bfq4htm6g3uypv7xawzyppipnt4"
},
"mimeType": "image/jpeg",
"size": 198284
},
"path": "/articles/d41586-026-01617-2",
"publishedAt": "2026-05-27T04:08:57.071Z",
"site": "https://www.nature.com",
"tags": [
"doi:10.1038/d41586-026-01617-2"
],
"textContent": "Nature, Published online: 26 May 2026; doi:10.1038/d41586-026-01617-2\n\nWhat happens when AI companies produce models that they say the public can’t have — and how should users and governments react?",
"title": "Too dangerous to release: is Mythos the start of the restricted-AI era?",
"updatedAt": "2026-05-26T00:00:00.000Z"
}