Turing Award Winner Rich Sutton Founds Oak Lab — AI Agents That Learn on Their Own
Rich Sutton, the father of reinforcement learning, launches Oak Lab to build AI agents that learn continuously from their environment in real time.
Rich Sutton's Vision for the Future of AI
Rich Sutton, the 2024 Turing Award winner widely regarded as the father of reinforcement learning, is launching a new venture. He and his former student Khurram Javed are founding Oak Lab, moving over from John Carmack's Keen Technologies. Sutton's mission: build AI agents that learn on their own, in real time, from direct interaction with their environment — no pre-training on static datasets required.
Why Sutton Calls Today's AI "Weak and Inefficient"
Sutton is blunt about the limitations of current AI. He calls deep learning methods "weak and inefficient," arguing they need "not more tweaks, but fundamentally new ideas and a thorough reworking before they can provide a solid foundation for achieving the more ambitious goals of AI." In June 2026, he argued that generative AI is good at imitation but can't evaluate its own outputs, making it incapable of real discovery.
Oak Lab's Different Approach
Instead of training on massive static datasets (as GPT-5.6 or Claude 4 do), Oak Lab's agents will learn by interacting with real environments, constructing internal world models, and using reinforcement learning — the technique Sutton pioneered. His recent statements describe a vision where AI agents learn continuously from their environment, handle variation, evaluation, and selection on their own.
The Ambitious Goal: 20 Watts, 1 Trillion Parameters
Oak Lab's long-term target is striking: "a trillion parameters that learns and plans in real time with 20 watts of energy." To put that in perspective, training a single GPT-5.6 class model consumes megawatts of electricity. Running it in real time at 20 watts — roughly the power consumption of a human brain — would represent a fundamental breakthrough in AI efficiency.
What Makes Oak Lab Different
Several factors distinguish Oak Lab: Sutton is one of the most cited and influential AI researchers alive — his methods are used across the entire AI industry; the team comes from Keen Technologies with deep reinforcement learning expertise; they operate in a less crowded space compared to the LLM race; and the 20-watt goal demands fundamentally new architectures, not incremental improvements.
Implications for the Georgian Ecosystem
Oak Lab's approach is particularly relevant for smaller markets like Georgia. If Sutton's vision succeeds, AI development may become less dependent on massive compute infrastructure and more efficient, making it accessible to more regions and businesses.
Next Steps
Oak Lab is still in its early stages, but Sutton's name and the Keen Technologies pedigree create serious expectations. While achieving their goals will take years, the direction — AI that learns autonomously, continuously, and efficiently — represents one of the most important trends in the future of artificial intelligence.