Google’s Chief Scientist, Jeff Dean, says there’s one hot topic in artificial intelligence he prefers to stay away from Artificial General Intelligence (AGI).
Speaking on “The Moonshot Podcast, Dean explained that AGI is too vague and means different things to different people. For some, it’s a realistic goal, while for others it’s almost impossible. Depending on who you ask, the difficulty of achieving AGI can vary by a factor of a trillion, Dean said, adding that this makes the debate confusing and unproductive.
Current State of AI
Dean believes today’s AI systems are already impressive. According to him, models now outperform the average person in many non-physical tasks. While they still can’t match human experts, he described them as “pretty reasonable at most things.”
This stands in contrast to the extreme idea of AGI a system that can beat humans at every possible task. Dean suggested that focusing on that definition distracts from the real, tangible progress being made.
The Future of AI
Looking ahead, Dean sees AI as a powerful tool for science and engineering. He expects automated search and computation to help researchers make discoveries faster in the coming decades.
“There will be a lot of domains where automated search and computation can accelerate progress scientific progress, engineering progress,” he said.
AGI Debate in the AI Community
The AGI debate remains heated. Some experts define it as human-like intelligence, while others see it as AI outperforming humans in most meaningful work. Predictions for when AGI might arrive vary widely from just a few years away to several decades in the future.
