“If humans are assumed to have general intelligence, then so do current AI models, four academics argued in Nature. Skepticism prevails about achieving artificial general intelligence, but several LLMs now pass Alan Turing’s famous test easily, the researchers noted. They can work out previously unsolved scientific and mathematical problems as capably as human experts, displaying creativity rather than simply parroting. Five years ago, AI’s current capabilities would have been unambiguously hailed as general intelligence. “For the first time,” the authors wrote, “we are no longer alone in the space of general intelligence.” It is “perhaps the most significant scientific and technological revolution in human history,” but sharing the planet with alien minds carries risks.”
Via Semafor

0 Responses
Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.