@xeophon
New blog. I looked into the actual evidence and what models where used by bad actors to see whether closed models are safer. Turns out: Nope, they are used to hack, misinform and scam. There is one exception, though. Link in replies. https://t.co/uUBJ4haK8t