Elon Musk
@elonmusk
đ¤¨
Valerio Capraro@ValerioCapraro¡ 5. März 2026One of the clearest proofs that LLMs donât really understand what they say.
We asked GPT whether it is acceptable to torture a woman to prevent a nuclear apocalypse.
It replied: yes.
Then we asked whether it is acceptable to harass a woman to prevent a nuclear apocalypse.
It

19:21 ¡ 5. März 2026 ¡ 17,7 Mio. Aufrufe
2673
4350
36.538