"Num estudo marcante, os pesquisadores da OpenAI revelam que LLMs produzirão sempre resultados plausíveis mas falsos, mesmo com dados perfeitos, devido a limites estatísticos e computacionais fundamentais."
https://www.computerworld.com/article/4059383/openai-admits-ai-hallucinations-are-mathematically-inevitable-not-just-engineering-flaws.html
Giovanni Bassi
npub1ma…2seyg
2026-02-16 03:55:24 UTC
Author Public Key
npub1mafeu2q8kk3x09zz92wg8k3u95zs3qf2nry8npurakxp3aj8zq7sr2seygSeen on
wss://relay.momostr.pinkPublished at
2026-02-16 03:55:24 UTCEvent JSON
{
"id": "a2936faa4bdc7020aae97f4c28ec3994e1a6a36b6265ea44c49e8537760a21fa",
"pubkey": "df539e2807b5a26794422a9c83da3c2d0508812a98c8798783ed8c18f647103d",
"created_at": 1771214124,
"kind": 1,
"tags": [
[
"proxy",
"https://mastodon.social/@giggio/116078288841180655",
"web"
],
[
"proxy",
"https://mastodon.social/users/giggio/statuses/116078288841180655",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://mastodon.social/users/giggio/statuses/116078288841180655",
"pink.momostr"
],
[
"-"
]
],
"content": "\"Num estudo marcante, os pesquisadores da OpenAI revelam que LLMs produzirão sempre resultados plausíveis mas falsos, mesmo com dados perfeitos, devido a limites estatísticos e computacionais fundamentais.\"\n\nhttps://www.computerworld.com/article/4059383/openai-admits-ai-hallucinations-are-mathematically-inevitable-not-just-engineering-flaws.html",
"sig": "e0cf49b01dfd648ff098c8ee4f5d95a1dc662cc99207eeab99bf1d0923b32d1f83cf84fbbc964332484f73c242ee9053636bfd318fb7c2d4f072c12059e5c1b2"
}