While chatbots have demonstrated impressive capabilities, DeepMind CEO Demis Hassabis cautions against overstating their current intelligence. In a recent appearance on the All-In Podcast, Hassabis argued that comparing them to possessing “PhD intelligences” is misleading, citing their inability to consistently perform basic reasoning tasks. He contrasted this with his vision of true Artificial General Intelligence (AGI), which he believes will exhibit robust continuous reasoning, adaptability, and learning without such glaring inconsistencies. Hassabis estimates that realizing genuine AGI is still 5 to 10 years away. The discussion was highlighted in a Reddit post by user /u/Nunki08.
DeepMind’s Hassabis: AGI Still a Distant Horizon Despite Chatbot Advances
