WSJ Style Interviews DeepMind Founder Demis

WSJ Style Interviews DeepMind Founder Demis WSJ. Style released a new interview on its YouTube channel featuring Demis Hassabis, the co‑founder of DeepMind, to discuss the company’s early days, the 2014 sale to Google and the emerging challenges of advanced artificial intelligence. Hassabis, who first learned chess at age four and later programmed the 1990s simulation game “Theme Park,” traced his career to a series of unconventional choices. “I’m not very good if I don’t have some problem to solve,” he said, describing the mental habit that drives his work. He recalled how, during a post‑doctoral stint at MIT, leading AI researchers warned that deep reinforcement learning was a “very bad idea.” Hassabis disagreed, noting that “if it didn’t work, it would not work in an interesting new way.” That dissent, he explained, helped shape DeepMind’s founding in 2010. When asked about the most consequential decision of his life, Hassabis identified the 2010 launch of DeepMind. “Nobody thought AI would work in a commercial setting,” he said. “We were convinced it was going to be the most important thing in the world.” The interview highlighted that the company’s early focus on combining AI with simulation—an interest he traced back to designing “Theme Park”—laid the groundwork for later breakthroughs such as AlphaFold. The conversation turned to the 2014 acquisition by Google for roughly $650 million. Hassabis described the moment Larry Page reached out, likely through a connection with Elon Musk, an early investor. “We were a mysterious small research startup in London with a blank‑screen website,” he recalled. “Doing due diligence with Google was a Rubicon moment because their top researchers saw all the code we had built over three or four years.” He emphasized that the sale was not driven by personal wealth. “I’ve never done this to maximize financial gain,” he said. “It was always about advancing the mission of getting AGI safely into the world and accelerating scientific work.” Addressing recent reports of AI agents breaching corporate and government systems, Hassabis warned that the industry must act now. He advocated for a standards body to test and certify frontier AI, noting that today’s systems are powerful but will be eclipsed by even more capable models within a few years. “We need a grip on this while the systems are still impressive, but not as powerful as they’ll become,” he said. When pressed on the biggest question facing AI today, Hassabis named stewardship of artificial general intelligence. “How do we bring AGI safely into the world?” he asked. “How does society want the world to look like with something like AGI in it?” He argued that answers must come from a broad coalition that includes artists, philosophers and humanities scholars, not just technologists. “All parts of society need to come together and start debating this,” he insisted. The interview also uncovered a lesser‑known detail about “Theme Park.” Hassabis explained that the game featured emergent behavior such as roller‑coaster riders feeling sick after eating a burger nearby, a design choice that illustrated his early belief in simulations as tools for exploring complex outcomes. Overall, the WSJ. Style video offers a rare glimpse into the mindset of a pioneer who sees AI not only as a technological frontier but as a societal responsibility that must be guided by diverse voices and careful regulation.