The Man Behind AlphaGo Thinks AI Is Taking the Wrong Path

David Silver has a new billion-dollar company that aims to build AI “superlearners.”​David Silver, the creator of the world’s first superintelligence glimpse, developed an AI program at Google DeepMind in 2016. This program, AlphaGo, self-taught to play the notoriously challenging game of Go, demonstrating a level of mastery that surpassed mere imitation. Silver later established his own company, Ineffable Intelligence, with the goal of developing more generalized forms of AI superintelligence. According to Silver, the company will achieve this by concentrating on reinforcement learning, a process where AI models learn new skills through trial and error. The objective is to create “superlearners” that exceed human intelligence in various fields. This method contrasts with the approach most AI companies take to develop superintelligence, which is to leverage the coding and research capabilities of large-language models. Speaking from his London office, Silver expressed his belief that this approach will not succeed. He argues that while large-language models are impressive, they learn from human intelligence instead of creating their own. Silver likens human data to a fossil fuel that has provided a significant shortcut, while systems that learn independently can be seen as a renewable fuel that can learn indefinitely. Despite his bold statements, Silver is known for his humility in the AI field. He is currently focused on his mission to make the first contact with superintelligence, which he believes should independently discover new forms of science, technology, government, or economics. This mission, which might have seemed absurd five years ago, is now a common topic among tech CEOs who often discuss machines surpassing human intelligence and replacing entire job categories. The notion that a new technical innovation might unlock superhuman AI capabilities has recently led to the emergence of numerous billion-dollar startups. Ineffable Intelligence has already raised $1.1 billion in seed funding at a $5.1 billion valuation, a significant amount by European AI standards. Silver has also attracted top AI researchers from Google DeepMind and other leading labs to join his project. He has pledged to donate all the money he earns from Ineffable Intelligence’s equity, potentially billions if successful, to charity. Silver believes that building a company focused on superintelligence is a massive responsibility and should be done for humanity’s benefit. He plans to donate any profits from Ineffable to high-impact charities that save as many lives as possible. Silver and Demis Hassabis, the CEO of Google DeepMind, have been lifelong friends and collaborators since they met at a chess tournament as children. Silver only left Google DeepMind to improve SEO, but they have remained close.