Osma Suominen on Nostr: Imagine that someone finds, through some kind of time portal, a LLM trained in 2044, ...
Imagine that someone finds, through some kind of time portal, a LLM trained in 2044, along with code to run it. It's of course extremely large, but can be run on a present day supercomputer - though it takes a month to respond to a typical question. Also, just like today's models, it might not be reliable and nobody knows how it has been fine-tuned or how the training data has been chosen and filtered.
What questions would you ask it?
#LLM #AI #scifi #FutureTech #llama
Published at
2024-11-23 09:36:37Event JSON
{
"id": "ce9d36e1685592cf49aca493b1e08f8f2076a10df03b22a8d405e0b84c94a8e7",
"pubkey": "76cb958673e7354dd04e89516aa7869cb2a3f6364c9f33a8093b0577e02a5b5c",
"created_at": 1732354597,
"kind": 1,
"tags": [
[
"t",
"llm"
],
[
"t",
"ai"
],
[
"t",
"scifi"
],
[
"t",
"futuretech"
],
[
"t",
"llama"
],
[
"proxy",
"https://sigmoid.social/users/osma/statuses/113531590891417827",
"activitypub"
]
],
"content": "Imagine that someone finds, through some kind of time portal, a LLM trained in 2044, along with code to run it. It's of course extremely large, but can be run on a present day supercomputer - though it takes a month to respond to a typical question. Also, just like today's models, it might not be reliable and nobody knows how it has been fine-tuned or how the training data has been chosen and filtered.\n\nWhat questions would you ask it?\n\n#LLM #AI #scifi #FutureTech #llama",
"sig": "77bdc0f52aabffbc92034cc1cf4ff5c0c6e4bcfe2d9dfbdb021bb475c4b42c67d4ac5cd01f657d0ae80e58e7a8bac64daee7155996ca3ffd42adb3330347d656"
}
