😀
I know very well (and faced several times) curve overfitting: it's a term related to models working very well on the "training data", but badly on new samples outside that set.
Using the term for LLM output reproducing verbatim fragments of the training data is just mischaracterizing the term to fool those who do not understand statistical algorithms (and neural networks in particular).
As for Shannon, you seem to ignore what lossy compression is: https://en.wikipedia.org/wiki/Lossy_compression
Having said that, I'm afraid that no amount of computer science or examples can change the mind of an #AI believer/addict.
Giacomo Tesio
npub1ng…qgmse
2026-08-23 17:59:56 UTC
in reply to nevent1q…lm6a
Author Public Key
npub1ngr980mlhd72qa6qw0guv6cz4pccqrz28lx2rxflh6xcchfgl7ysaqgmseSeen on
wss://relay.momostr.pinkPublished at
2026-08-23 17:59:56 UTCEvent JSON
{
"id": "17d942d593d7d802329d0d5bde5d4d8a579b3caa169ff1e61ee4b9b16018c2d8",
"pubkey": "9a0653bf7fbb7ca0774073d1c66b02a871800c4a3fcca1993fbe8d8c5d28ff89",
"created_at": 1787507996,
"kind": 1,
"tags": [
[
"proxy",
"https://snac.tesio.it/giacomo/p/1787507996.583349",
"web"
],
[
"p",
"a19807f96b90dfa8fec92966cde944f55c54a2d704337417449c801dde3ee72c"
],
[
"e",
"d01ce5b7fc8e7a8c07b881fa11b15998abb683e2a3d97640df49916c0f2f221c",
"",
"reply",
"2aad90c13120482441b028cb1eefad12fafca60fa257c35fe7aedf5c94c6b90a"
],
[
"t",
"ai"
],
[
"p",
"2aad90c13120482441b028cb1eefad12fafca60fa257c35fe7aedf5c94c6b90a"
],
[
"p",
"9a0653bf7fbb7ca0774073d1c66b02a871800c4a3fcca1993fbe8d8c5d28ff89"
],
[
"e",
"62e8b71eba6f2a19c2caf83ce9b57ff5392d4be35fbc9ce85fa0b2d3c24c34e8",
"",
"root",
"a19807f96b90dfa8fec92966cde944f55c54a2d704337417449c801dde3ee72c"
],
[
"proxy",
"https://snac.tesio.it/giacomo/p/1787507996.583349",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://snac.tesio.it/giacomo/p/1787507996.583349",
"pink.momostr"
],
[
"-"
]
],
"content": "😀\n\nI know very well (and faced several times) curve overfitting: it's a term related to models working very well on the \"training data\", but badly on new samples outside that set.\n\nUsing the term for LLM output reproducing verbatim fragments of the training data is just mischaracterizing the term to fool those who do not understand statistical algorithms (and neural networks in particular).\n\nAs for Shannon, you seem to ignore what lossy compression is: https://en.wikipedia.org/wiki/Lossy_compression\n\nHaving said that, I'm afraid that no amount of computer science or examples can change the mind of an #AI believer/addict.",
"sig": "41026a5f4ee7464597b745165610e6642607c304859789b5b8b93eb46dee1c6ee5d9d7df2b47832afde5d2d16deb1c2c0b73df2023d9d0cc5c5908f23594f789"
}