Skyfall AI, co-founded by the creators of Maluuba, has launched a new initiative aimed at developing the first Autonomous Enterprise by advancing beyond traditional large language models (LLMs). They introduced Morpheus, a business simulation that allows for the compounding effects of past decisions, contrasting with standard reinforcement learning benchmarks that reset after each episode. This approach is part of Skyfall AI’s larger goal to implement Enterprise World Models, which simulate the consequences of strategic actions. To support this vision, the lab plans to acquire small SaaS startups valued at up to $1 million, which will serve as real-world applications for their innovative models, addressing the growing demand for more sophisticated AI solutions that function effectively in dynamic environments.

Maluuba: Maluuba was an early AI research lab established to build foundational modules for deep learning and reinforcement learning. Acquired by Microsoft, it became part of the company’s Canadian research efforts and collaborated with leading academics in the field. Its original mission of advancing core AI capabilities continues to influence the vision behind Skyfall AI.
Microsoft: Microsoft is a major technology company with extensive AI research operations, including MSFT Research Canada. It acquired Maluuba, integrating the lab’s pioneering work in deep learning and reinforcement learning into its broader research portfolio. This history connects directly to the founders’ current efforts at Skyfall AI to push beyond established AI methods.
Skyfall AI: Skyfall AI is a frontier neo lab founded by veterans of early AI research to develop next-generation models beyond current LLM paradigms. It focuses on Enterprise World Models using continual learning and world modeling to enable autonomous enterprises that simulate and act on complex, evolving business environments. The lab recently launched the Morpheus platform and is pursuing acquisitions of small SaaS businesses to advance and validate its approach.
Victor Dey: Victor Dey is a journalist who conducted an in-depth interview with Skyfall AI’s founders for a recent Forbes feature. His coverage explores the lab’s long-term vision for autonomous enterprises and next-generation AI models. The article provides external context on the team’s transition from Maluuba to Skyfall AI’s current initiatives.
Yoshua Bengio: Yoshua Bengio is a Turing Award-winning researcher known for foundational contributions to deep learning. He collaborated with the Maluuba team on early scientific AI projects during a period of pure research before widespread hype. This background aligns with Skyfall AI’s focus on rigorous, paradigm-shifting advancements in frontier models.
Kaheer Suleman: Kaheer Suleman is a co-founder of Maluuba and serves as chief AI officer at Skyfall AI, drawing on decades of experience in foundational AI research. He is driving the development of Enterprise World Models to address long-horizon planning and adaptation challenges that current LLMs struggle with. His work emphasizes first-principles approaches to continual learning in dynamic real-world settings.
Richard Sutton: Richard Sutton is a Turing Award-winning expert in reinforcement learning who contributed to Maluuba’s foundational work on core AI modules. His principles of RL inform efforts to create persistent, non-stationary learning systems at new labs like Skyfall AI. The connection highlights the evolution from episodic benchmarks to continual, real-world applications.

AI Paradigm Shift: Frontier AI efforts are shifting toward world models that can simulate cause-and-effect outcomes to overcome limitations of static LLM approaches in complex settings.
Research Innovation: Skyfall AI recently introduced Morpheus, a persistent enterprise simulation benchmark designed to advance continual reinforcement learning in evolving, non-stationary environments.
Enterprise Automation: Skyfall AI is acquiring small software businesses to fully automate their operations, using them as real-world testbeds for developing autonomous enterprise systems.