Jukka Niiranen on Nostr: As #OpenAI teaches its new model to "reason", it also becomes a better, more ...
Published at
2024-09-18 17:38:29Event JSON
{
"id": "b41baac9a76c730042cabfd3e45f9820ed6d18c09c4c92680f24228e8e6cec94",
"pubkey": "158d3103117c5c776ea82b4661e30480de444a998d9bb2d048ac19b1c4f322ca",
"created_at": 1726681109,
"kind": 1,
"tags": [
[
"t",
"openai"
],
[
"proxy",
"https://mstdn.social/users/jukkan/statuses/113159773197065161",
"activitypub"
]
],
"content": "As #OpenAI teaches its new model to \"reason\", it also becomes a better, more intentional and deceptive liar. No biggie! https://www.theverge.com/2024/9/17/24243884/openai-o1-model-research-safety-alignment\n\nhttps://media.mstdn.social/media_attachments/files/113/159/770/177/924/427/original/e83be2af2c7e3d91.png",
"sig": "b50c8cfbeee93f65725b733d8cd57e736105247261d5024dfaa8765c607a78eb5bc73eeb41dc45da575eeb1871d00210d66c682deb58b02efbd63fc8c0b70708"
}
