Lucky ✨"i prompted an output from the machine that outputs crazy bullshit with no basis in reality and what happened next will shock you"
grandpa no!
Brint@lucky I took Bernie's video to be a display of how the AI "works" and felt pretty well done. At one point it even disagrees with his statements on regulating it, and he easily baits it into agreeing with him in the next response. It puts Anthropic in the position of either agreeing with the statements that their AI is dangerous and should be regulated, or coming out and saying "yeah, everything it said is bullshit".
I feel like it doesn't do much for those of us who understand the probabilistic nature of these LLMs (other than maybe make us chuckle and cringe). But for those who don't understand it, it should either scare them into understanding the dangers, or make them doubt the validity of the output... both of which feel like a positive outcome to me.
Having said all this, I did just read through this perspective and between your comments and this, it has me rethinking my position.
To me, it felt like an attack on Claude, not an ad for it... But maybe that's only because I'm already on that side.
https://dair-community.social/@timnitGebru/116263058466613099
Timnit Gebru (she/her). (@[email protected])
Bernie Sanders just created a video to make a point about issues of privacy in AI (a real problem). His supporters on Muskrat's platform came for me when I quote tweeted his post saying, "at this point, Bernie’s job seems to be lobbying for Anthropic." So I'm going to elaborate on the corporate psyop that has been happening since the beginning of time and how no one seems to learn their lesson.