TFTC on Nostr: The White House is preparing to bring open-source AI models under its prerelease ...
The White House is preparing to bring open-source AI models under its prerelease safety testing framework.
Once open models reach frontier capabilities, they'd face a potential 30-day government testing period before public release.
The catch: officials know that excluding open models creates a government-approved advantage for closed labs like OpenAI and Anthropic. But including them could slow US open-model development and hand China an edge.
The framework is technically voluntary for now, but the competitive pressure to get a government seal of approval could make it mandatory in practice.
Published at
2026-08-12 23:22:24 UTCEvent JSON
{
"id": "065018ee473a249d8766930a9ca3a45a137e54ca3121ead89bb7e273816b3e94",
"pubkey": "85bdb5875e113dcc99498b474636449882ee15a1c78154ab07c065b47339d672",
"created_at": 1786576944,
"kind": 1,
"tags": [
[
"imeta",
"url https://blossom.primal.net/ad774bd1e959a987b22880b9602d6e063fae9b7973fd720efb8dd3b8aaf4a11d.jpg",
"m image/jpeg",
"x ad774bd1e959a987b22880b9602d6e063fae9b7973fd720efb8dd3b8aaf4a11d",
"size 119389"
]
],
"content": "The White House is preparing to bring open-source AI models under its prerelease safety testing framework.\n\nOnce open models reach frontier capabilities, they'd face a potential 30-day government testing period before public release.\n\nThe catch: officials know that excluding open models creates a government-approved advantage for closed labs like OpenAI and Anthropic. But including them could slow US open-model development and hand China an edge.\n\nThe framework is technically voluntary for now, but the competitive pressure to get a government seal of approval could make it mandatory in practice.\nhttps://blossom.primal.net/ad774bd1e959a987b22880b9602d6e063fae9b7973fd720efb8dd3b8aaf4a11d.jpg",
"sig": "3b09d9ac08128912ca9b99e3e30079e0bb8b20a4fe6a23dc219df690b93dbc87317175dd65490ad8ce47851599b984015c277044f17e74d52ec554387c0df9bd"
}
