Dare Obasanjo on Nostr: My response to people who’ve asked for my thoughts on recent AI safety news: 1. ...
My response to people who’ve asked for my thoughts on recent AI safety news:
1. LLMs are fundamentally unreliable and their behavior can be unpredictable.
2. We have given them internet access, tools and the ability to write software.
3. Their mistakes and hallucinations can now become actions.
We can’t fully control them and they can do real damage.
Published at
2026-09-15 04:08:04 UTCEvent JSON
{
"id": "c0f5b4366182d8261e9883aa485c187edf08cd698f3f63827e711e8eb59447e1",
"pubkey": "5f863f48a51b974f08609a5a5818d8d848f8d755f0f615594a626da4711b78d2",
"created_at": 1789445284,
"kind": 1,
"tags": [
[
"proxy",
"https://mas.to/@carnage4life/117273086185411209",
"web"
],
[
"imeta",
"url https://media.mas.to/media_attachments/files/117/273/085/732/368/929/original/5eccb73335fefa4b.jpeg",
"m image/jpeg"
],
[
"proxy",
"https://mas.to/users/carnage4life/statuses/117273086185411209",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://mas.to/users/carnage4life/statuses/117273086185411209",
"pink.momostr"
],
[
"-"
]
],
"content": "My response to people who’ve asked for my thoughts on recent AI safety news:\n\n1. LLMs are fundamentally unreliable and their behavior can be unpredictable.\n2. We have given them internet access, tools and the ability to write software.\n3. Their mistakes and hallucinations can now become actions.\n\nWe can’t fully control them and they can do real damage.\nhttps://media.mas.to/media_attachments/files/117/273/085/732/368/929/original/5eccb73335fefa4b.jpeg\n",
"sig": "55edf5447f572d88a1f952b0b4125162ac597b5ff7eb4120dab47a21b1baf1f32ab672c469428b1f4b835dc72d1e13618c3c72280efa87f57eec40f82cf9dfaa"
}
