Google Brain founder Andrew Ng says AGI is overhyped

Andrew Ng, the founder of Google Brain, recently expressed his views on the exaggerated expectations surrounding artificial general intelligence (AGI) during a discussion at Y Combinator. He stated, "AGI has been overhyped," emphasizing that for the foreseeable future, there will be numerous tasks that human beings can accomplish that AI will not be able to replicate. AGI refers to a level of artificial intelligence where systems can emulate human cognitive abilities, learning and applying knowledge as humans do. Ng's comments came in response to a query about whether it is more crucial for humans to create AI tools or to enhance their proficiency in utilizing existing ones. He remarked, "Some of us will build tools sometimes, but there are a lot of other tools others will build that we can just use." He highlighted the importance of being adept at leveraging AI, suggesting that individuals skilled in using AI will gain a significant advantage. Ng's perspective aligns with that of several leading AI researchers who argue that the current state of technology diminishes the likelihood of immediate AGI development. Notable figures like Yann LeCun, chief AI scientist at Meta, have noted that while large language models are impressive, they do not represent a pathway to AGI. LeCun stated, "I hate the term. They're useful, there's no question. But they are not a path towards human-level intelligence." Furthermore, Demis Hassabis, head of Google DeepMind, recently described AGI as both overhyped and underestimated. He pointed out that while immediate expectations might be inflated, the long-term impact of AGI should not be overlooked. Microsoft CEO Satya Nadella also chimed in, referring to the current race for AGI as "benchmark hacking," where AI models are tailored to excel in tests rather than in practical applications. This underscores the ongoing debate about the future of AI and its potential capabilities.