Turing Award winner Richard Sutton’s new venture is a bet against today’s AI models | Technology News

3 min readNew DelhiJul 14, 2026 01:51 PM Richard Sutton, one of the leading figures in the world of artificial intelligence, has announced the formation of a new AI lab built around the idea of AI agents that learn continuously from their environment.
Based in Toronto, Canada, Oak Lab has been founded Sutton and his former student Khurram Javed. Both of them previously worked at Keen Technologies, a startup pursuing artificial general intelligence (AGI) and located in Dallas, Texas, US, that was founded legendary gaming developer John Carmack.
With his new venture, Sutton is betting that reinforcement learning will define the next phase of AI. He argues that progress depends on moving beyond large language models (LLMs) trained on static datasets toward AI systems that learn from experience. The AI researcher has previously described current deep learning methods as weak and inefficient.

Oak Lab will be focused on developing algorithms that “learn in real time without storing or replaying data,” and constantly improve while using less processing power. The firm will also work on developing internal ‘world’ models while handling variation, evaluation, and selection on their own.
The long-term goal of Oak Lab is to build an agent with “a trillion parameters that learns and plans in real time with 20 watts of energy.” This new agent, called Oak, will also  keep evolving with experience.
Who is Richard Sutton?
Sutton is known as the co-founder of modern reinforcement learning, a technique that is vital for training LLMs powering viral AI chatbot such as Claude or ChatGPT. Simply put, reinforcement learning is a way for AI systems to learn from the digital equivalent of pleasure and pain.
In 2024, Sutton was awarded the Turing Award, the Nobel Prize-equivalent of computing, along with another researcher named Andrew Barto for their joint work on reinforcement learning.Story continues below this ad
Sutton is also a professor of computing science at The University of Alberta in Edmonton, Canada. After Google’s DeepMind unit closed the Edmonton lab he helped lead, Sutton joined Keen Technologies in September 2023.
Keen Technologies is backed several high-profile investors, including Shopify CEO Tobi Lütke. Its main goal is to build an artificial general intelligence (AGI) system with human-matching or -beating capabilities. However, Sutton said that he and Javed quit Keen Technologies because the pair wanted to take “a slightly different path” that involves rethinking deep learning entirely.

More importantly, Sutton is part of an increasingly vocal group of researchers, including AI pioneer Yann LeCun, who argue that present-day LLMs are a dead-end. They believe that the scaling laws that once led to sharp gains in model performance are broken, and that simply throwing more GPUs (Graphics Processing Units) at the models will not lead to more accurate, higher-quality responses.
In June this year, Sutton said that generative AI cannot be capable of real discovery as the underlying models are only good at imitating their training data and are not equipped to evaluate their own outputs.