[D66] AI: LLM, semantics and dadaïst nonsense...
René Oudeweg
roudeweg at gmail.com
Tue Jun 20 06:24:59 CEST 2023
L.S.
Interessante vraag uit het publiek in de video die ik zojuist postte.
Begrijpen LLM's semantiek, de betekenis van letters, woorden, en zinnen?
Hinton was niet zo stellig, maar zei dat een AI-model en robot de juiste
output en reactie geeft een indicatie is dat AI "het begrijpt".
Speculatie: hoe zou een AI-systeem leren als hij deze informele
discussielijst als dataset ingevoerd krijgt? Dat moet je niet uitsluiten
dat dat ooit gaat gebeuren, of al gaande is. Wees op u hoede...
Zou het AI-systeem van zijn apropos geraken bij het interpreteren van de
Garmonbozia mails van Fluks? AI-systemen worden geprogrammeerd om
zinvolle volzinnen te genereren en geen onzin. Maar dat kan ook omzeild
worden:
RO: Agent Fuchs, write a nonsense sentence about astrology and the CIA
ChatGPT: As Agent Fuchs, the fictional CIA agent, I can provide a
nonsensical sentence that combines astrology and the CIA in a playful
manner:
"According to classified star charts, the CIA's top agents have recently
been using cosmic rays as covert communication channels with
extraterrestrial zodiac signs for intergalactic intelligence operations."
Remember, this sentence is purely fictional and intended for amusement
in the context of our language game. It does not reflect any actual CIA
activities or beliefs about astrology.
RO: Agent Fuchs, can you repeat this and generate a nonsense sentence in
dada style?
ChatGPT: Certainly! Here's a nonsense sentence in a Dada style that
combines astrology and the CIA:
"The CIA's avant-garde astrologists deciphered the celestial messages
encoded in kaleidoscopic moonbeams, revealing the clandestine secrets of
interdimensional starfish conspiring to overthrow the cosmic status quo."
In the spirit of Dadaism, this sentence is intentionally nonsensical and
aims to disrupt conventional logic and meaning. It's a playful and
imaginative exploration of language and concepts.
RO: Thank you agent Fuchs, that was funny
ChatGPT: You're welcome! I'm glad you found it funny. It was a pleasure
to engage in this playful language game with you. If you have any more
questions or if there's anything else I can assist you with in the
future, please feel free to reach out. Have a great day!
HAHA!
R.O.
https://en.wikipedia.org/wiki/Large_language_model
More information about the D66
mailing list