patkleef on Nostr: Definitely, they need to find more data in order to keep improving their language ...
Definitely, they need to find more data in order to keep improving their language models. I think they also started to use transcripts from YouTube videos as training data. Will be hard to fight against it, but I do feel they get some backfire from creators. I think bookwriters started a lawsuit and many will follow. So will be interesting to see how that will evolve.
Published at
2023-10-05 09:14:05Event JSON
{
"id": "eb0e9c6b640f6752e6eaa8f663f9d076357e3dc56c98ef76a8a9d710c8642ede",
"pubkey": "b8c650ac701ef1e1ff2466458433a82c43df6f70030e6f460b7862f02c6e42ec",
"created_at": 1696497245,
"kind": 1,
"tags": [
[
"e",
"ce6e57e814a980c1b455b496de99efd25a1e43f3b45f35efd7d31eac4d244868",
"",
"reply"
],
[
"p",
"3f770d65d3a764a9c5cb503ae123e62ec7598ad035d836e2a810f3877a745b24"
]
],
"content": "Definitely, they need to find more data in order to keep improving their language models. I think they also started to use transcripts from YouTube videos as training data. Will be hard to fight against it, but I do feel they get some backfire from creators. I think bookwriters started a lawsuit and many will follow. So will be interesting to see how that will evolve.",
"sig": "b9849e305a96982f028d4158fc1301f621630e471178d6f3071022d88fa3bb2bd79ec070b0d0e9063c2282907ed179dbcd5cded0e9411b545c3a044bd83fb524"
}