This... is extremely interesting.
The point that I think a lot of folks are missing is that if you prompt AI with questions about the dangers of AI, the data it is trained on will admit that AI is dangerous.
add a skeleton here at some point
7 days ago