tech
We need to regard text as a deepfake medium.
I enjoyed this piece from Ted Chiang in The Atlantic because it’s a well-deserved smack in the face to the clowns who suggest LLMs might be conscious. Nope, we’re just falling for one of the most obvious confidence tricks of all time: giving in to “someone” who appears to like us.
TL;DR
- The author believes the idea of LLM consciousness is a confidence trick.
- People are falling for the AI's apparent friendliness.
- The suggestion of LLM consciousness deserves strong criticism.