Show Notes
This meeting featured an interview between Patrick, a professor managing the Master in Financial Engineering program at Lehigh University, and Nadab (also known as Nitty), founder and managing partner of Excite Capital. The discussion centered on Excite Capital's proprietary high-frequency trading platform, which leverages physics-based artificial intelligence rooted in quantum mechanics. The conversation explored how this approach fundamentally differs from conventional quantitative trading strategies by prioritizing real-time navigation over historical pattern recognition.
Key Concepts or Theories
Prediction vs. Navigation: Traditional AI recognizes patterns from historical data to forecast outcomes; Excite Capital's system instead navigates in real time by understanding the current state of the market and determining what moves remain possible. Observation, State, and Constraint: Three foundational concepts underpinning the firm's framework — what is visible (observation), what is actually occurring beneath the surface (state), and what limits the system's possible directions (constraint). Quasi-Particle Modeling: Individual stocks are modeled similarly to quasi-particles in physics — entities whose behavior only makes sense in relation to the larger field (market) surrounding them. Data Hamiltonian: A mathematical algorithm adapted from classical physics that represents both numeric and symbolic data in a single function, describing how a system's state evolves over time. Absolute, Hard, and Soft Rules: A three-tiered rules framework governing the autonomous trading platform, ranging from inviolable risk limits to dynamic, judgment-based parameters updated through reinforcement learning. Quantum Mechanics as a Framework: Used not to suggest stocks are quantum particles, but to provide a language for understanding real-time information, state evolution, possibility, interaction, and uncertainty in complex systems. Nature as Efficiency Model: Nature's ability to navigate an uncertain future by narrowing possibilities through constraints and updating continuously serves as a conceptual inspiration for Excite Capital's AI design. Important Questions Raised
How does navigation-based AI differ fundamentally from prediction-based AI in the context of financial markets? What role do quantum mechanics and the Hamiltonian model play in understanding market state evolution? How do natural constraints — such as capital, liquidity, and risk limits — reduce the universe of possible market outcomes? How does the human brain's instinctive, cost-efficient decision-making parallel the design of Excite Capital's AI system? Can physics-based AI models outperform conventional quant strategies that rely on speed, historical data, or headline reaction?
More from The Quant / Financial Engineering Podcast
It was my privilege to speak to Frank R. Gunter, a man of experience, wit and perspective on the Middle East based on his experience as a Marine, economist and having spent time in Iraq.