London-based David Silver led work on reinforcement learning (RL) at the Google AI unit. He founded his new firm, Ineffable Intelligence, in November. (Fortune)

Talking point: Silver is one of the most influential products of the Alberta school of AI. The British computer scientist got his doctorate at the University of Alberta under RL pioneer Richard Sutton. Silver was among the first of what would become a significant stream of Edmonton-trained researchers who worked for DeepMind in London. He took with him a project to get an RL system to play the strategy game Go. That work formed the basis of AlphaGo, a DeepMind model that beat human champion Lee Sedol four games to one in an exhibition series in March 2016—a breakout moment for the firm. Ineffable Intelligence is one of a number of so-called neolabs started by key researchers leaving larger tech companies. Others, like Ilya Sutskever’s Safe Superintelligence, Jerry Tworek’s Core Automation and Mira Murati’s Thinking Machines Lab, have attracted significant funding as the field chases more advanced AI.