They helped pioneer deep learning in Canada. Now they're building AI to replace CEOs
Essential brief
Former Maluuba co-founders Sam Pasupalak and Kaheer Suleman have launched Skyfall AI, a company developing AI world models designed to simulate the physical world and autonomously manage business o
Key topics
Key facts
Highlights
Why it matters
Skyfall AI's development of world models to autonomously manage company operations represents a significant evolution in AI applications beyond language processing. By aiming to replace human CEOs with AI systems, this approach challenges traditional business management structures and raises important considerations about the future of work and employment. The success of such technology could reshape industries by automating complex decision-making and operational tasks at scale.
Maluuba, a deep learning research lab based in Montréal and Kitchener-Waterloo, was a leading force in Canada's AI ecosystem before its acquisition by Microsoft in 2017. The lab's early work on deep learning laid the foundation for many of the large language models (LLMs) that drive today's AI boom. In 2026, two of Maluuba's co-founders, Sam Pasupalak and Kaheer Suleman, have returned with a new venture called Skyfall AI, headquartered in Toronto and San Francisco. Their focus is on developing AI world models, which simulate the physical world to perform complex tasks autonomously.
Skyfall AI aims to demonstrate that these world models can manage all operational aspects of a company with little or no human intervention. The company recently announced plans to acquire an existing micro-B2B-SaaS business for up to $1 million USD and replace its CEO with an AI system. Supported by investors such as Inovia Capital and Garage Capital, Skyfall's goal is to double the acquired company's revenue within six months through autonomous management.
Unlike LLMs trained primarily on text data, world models are trained on physical, spatial, and movement data, creating an internal 3D representation of the environment. This approach allows the AI to understand cause and effect rather than relying solely on pattern recognition. This technology has gained traction recently, with startups like General Intuition valued at $2.3 billion USD for their work on world models applied to robotics and physical AI.
Skyfall's approach represents a departure from Maluuba's original mission to develop conversational AI focused on natural language understanding. Pasupalak emphasizes real-world applicability and long-term planning capabilities of world models, which he believes surpass LLMs in handling high-stakes decision-making under uncertainty. The company plans to publicly document its progress in automating company operations.
While Skyfall acknowledges that certain human skills, such as managing interpersonal relationships, cannot be replicated by AI, the company expects to reduce the need for human operational roles. Employees of acquired companies will be offered generous departure packages and job placement support. This model raises questions about the impact of AI on employment, aligning with broader concerns about job displacement due to AI integration.
Pasupalak envisions AI handling monotonous operational tasks, freeing humans to focus on creative and passionate work. He compares this shift to historical technological advances like the printing press, which transformed labor rather than simply eliminating jobs. If successful, Skyfall aims to expand its technology to automate all parts of businesses globally, targeting a market potentially worth trillions of dollars.
Key topics in this update include they helped pioneer deep learning, they helped pioneer, and deep learning.