Experts Debate Astra's AGI Potential

Martin Ford posted that without human-like continual learning capability over long periods, Astra would not qualify as AGI. He argued that even a system passing Demis Hassabis's proposed test of independently discovering general relativity would remain a jagged intelligence without on-the-job learning capability. Gary Marcus retweeted Ford's remarks. Separately, 𝒻 (x) 🇲🇾 quoted Marcus's remark that as a scientist, it’s disappointing that we don’t (yet) know much about how the system actually works.