Mastodon Feed
@brib @grimalkina this kind of weaponization of the tools does leave open a really important question: is the existence of such activities evidence of use an “abuse” of LLMs, making them merely a tool so powerful that they can only be used for good or for evil, or (despite the various positive data points) is there something structural within the technology itself that lends itself to negative outcomes? I have my own answer but it’s definitely not trivial or obvious.