We could use "Word Vectors" to sort that out in a fairly-objective way.
If the correlation between "good father" and "trusted to look after my cat" is 0.9, then that's the weighting that should apply if the user's intent is to find someone to look after her cat.
Word Vectors (e.g. from Spacy or NLTK) also allow clients to handle arbitrary attestation strings.
Sentence Transformers could also do this, but are much less performant
Raison d'État
npub1mt…3ueyt
2026-02-26 21:54:36 CET
in reply to nevent1q…nar0
Author Public Key
npub1mts2l7rsvgpqxdc8sk39e50fpntw0p7eamq8pp3kek5h6k7qemzq83ueytPublished at
2026-02-26 21:54:36 CETEvent JSON
{
"id": "09e6e807693050cf2b639535715fecea1a07b83cc87618bae2829469834ba36e",
"pubkey": "dae0aff870620203370785a25cd1e90cd6e787d9eec0708636cda97d5bc0cec4",
"created_at": 1772139276,
"kind": 1,
"tags": [
[
"alt",
"A short note: We could use \"Word Vectors\" to sort that out in a ..."
],
[
"e",
"9ac90bee2f9457595d07bfbc558865a202f67ac23bb1d26cb305dcf0b2d85bc4",
"wss://vitor.nostr1.com/",
"root",
"460c25e682fda7832b52d1f22d3d22b3176d972f60dcdc3212ed8c92ef85065c"
],
[
"e",
"8d91b821dd437f37f74548efb878c728df9b95e546372f7700c61bbffcc666ba",
"wss://relay.damus.io/",
"",
"c4f5e7a75a8ce3683d529cff06368439c529e5243c6b125ba68789198856cac7"
],
[
"e",
"3140232a32b3087066f151ffcc24a5ce8d567ef5b313b3326393e510ff9da0e2",
"wss://vitor.nostr1.com/",
"reply",
"460c25e682fda7832b52d1f22d3d22b3176d972f60dcdc3212ed8c92ef85065c"
],
[
"p",
"460c25e682fda7832b52d1f22d3d22b3176d972f60dcdc3212ed8c92ef85065c",
"wss://nostr.wine/"
],
[
"p",
"c4f5e7a75a8ce3683d529cff06368439c529e5243c6b125ba68789198856cac7",
"wss://relay.day.ag/"
]
],
"content": "We could use \"Word Vectors\" to sort that out in a fairly-objective way. \n\nIf the correlation between \"good father\" and \"trusted to look after my cat\" is 0.9, then that's the weighting that should apply if the user's intent is to find someone to look after her cat.\n\nWord Vectors (e.g. from Spacy or NLTK) also allow clients to handle arbitrary attestation strings.\n\nSentence Transformers could also do this, but are much less performant",
"sig": "982c3aa8ff350b00463dc54028a5c047bc67937e9d8d3b69719bced83fcec2d38808baaed58d269871833caa43407a7efc19ca3cc128ad8d1f457203ae87e3b7"
}