AsianFin – On the evening of November 23, NVIDIA founder Jensen Huang, who was recently awarded an honorary doctorate in engineering by the Hong Kong University of Science and Technology (HKUST), engaged in a wide-ranging discussion with Harry Shum, Chair of HKUST’s Council and a foreign member of the U.S. National Academy of Engineering. The conversation touched on topics such as digital intelligence ecosystems, scaling law, digital twins of human biology, physical intelligence, technological flywheels, university resource shortages, leadership, and energy consumption. Here are the 12 key points from Huang’s remarks:
1. Core of AI: Huang emphasized that the essence of artificial intelligence lies in automating cognition—solving problems through “digital agents.” These agents form what he calls “digital intelligence,” which, when hosted in a data center, functions like a factory producing intelligence.
2. Digital Intelligence Ecosystem: AI is building a new ecosystem of “digital intelligence,” which Huang described as an emerging industry that converts energy into intelligence for use across various applications. The demand for such intelligence will be enormous.
3. Scaling Law and Specialization: While scaling law remains effective, pre-training alone is insufficient. Huang stressed the importance of fine-tuning AI for specific fields through reinforcement learning, human feedback, AI feedback, synthetic data generation, and multi-path learning. These techniques allow AI to acquire complex and specialized skills.
4. NVIDIA’s Advances in Computing: Over the past decade, NVIDIA has reduced computing costs by a factor of one million, fundamentally changing how people approach computation. This ease of use has fueled the rapid growth of machine learning.
5. AI and Simulation: Unlike traditional solvers based on first principles, AI derives answers by learning from data, making it a powerful tool for simulation. Simulation enables researchers to explore complex systems on a larger scale, showcasing AI’s transformative potential for scientific discovery.
6. Digital Twins for Human Biology: Huang highlighted the possibility of using AI to simulate human biology, creating digital twins to model and understand complex, multi-scale systems. This innovation could revolutionize fields such as digital biology and climate science.
7. University Resource Shortages: Addressing the structural issue of limited computing resources in universities, Huang proposed systemic changes, such as building shared infrastructure to support all disciplines. While challenging, such shifts are essential for long-term progress.
8. AI and Energy Efficiency: AI has the potential to reduce inefficiencies in various systems, such as energy grids, where power is often overproduced yet insufficient during peak demand. Huang envisions AI optimizing energy usage, making it the best possible use of energy.
9. Physical Intelligence: Future AI must learn to interact with the physical world, a concept Huang calls “physical intelligence.” This is central to robotics and presents significant opportunities in the Greater Bay Area, which is uniquely positioned to advance both AI and hardware ecosystems.
10. Technological Flywheel: Huang identified three types of robots—cars, drones, and humanoid robots—that have achieved scalable production. Large-scale manufacturing drives R&D investment, leading to technological breakthroughs that further enhance production. This “flywheel effect” is a critical driver of industry growth, with the Greater Bay Area uniquely positioned to integrate AI and electromechanics.
11. Leadership Principles: As a leader, Huang believes in maintaining confidence while embracing uncertainty. Leaders should inspire strength in their teams but also acknowledge vulnerability. Most importantly, leadership should focus on serving others and aligning every decision with the team’s mission and success.
12. Transparent Decision-Making: Huang directly manages 60 executives and stresses the importance of transparency. He ensures collective reasoning and discussion guide decision-making, acting as the final summarizer to eliminate ambiguity and ensure alignment.
