From Toy Problems to Real Impact: AI’s Evolution and the Rise of Embodied Intelligence
Artificial intelligence has undergone a fundamental transformation—moving from solving artificial problems to tackling real-world challenges that matter. This shift represents more than technological progress; it changes how we think about computing power and its applications.
The Journey from Academic Exercises to Practical Solutions
In the early 1980s, AI researchers worked on what Dr. Wang calls “toy problems”—artificial challenges designed for academic study rather than real-world application. The technology was shaky, and the problems were disconnected from genuine human needs.
Today’s AI operates differently. Modern systems solve actual problems that affect people’s lives and businesses. This evolution from artificial to authentic challenges marks AI’s maturation from laboratory curiosity to practical tool.
The transformation parallels how increased computing power changes human thinking itself. When you have a bicycle, you think differently about traveling from Hong Kong to Shanghai than when you have a car. Give someone an airplane, and their approach changes completely. A rocket opens entirely new possibilities.
Computing power works the same way. A million-fold increase in computational capability doesn’t just make existing processes faster—it fundamentally alters how we approach problems and what we consider possible.
Beyond the AGI-ASI Classification Framework
The popular progression from AI to Artificial General Intelligence (AGI) to Artificial Super Intelligence (ASI) oversimplifies reality. These classifications create artificial boundaries in what is actually a continuous evolution.
AI development resembles human growth from kindergarten to PhD—a gradual expansion of capabilities without clear demarcation points. Children possess remarkable learning abilities, just as PhD students do. The difference lies in accumulated knowledge and refined skills, not fundamental cognitive architecture.
Rather than focusing on categorical distinctions, we should recognize AI as an evolving capability that grows more sophisticated over time. Each advancement builds on previous achievements without requiring revolutionary leaps between artificial categories.
The Convergence Toward Embodied AI
Three years ago, natural language processing, computer vision, and robotics operated as separate disciplines. Today, these fields merge into a unified AI ecosystem where advances in one area immediately benefit others.
This convergence enables embodied AI—artificial intelligence that operates in physical environments through robotic systems. Robotics provides the platform where AI capabilities deploy in the real world, similar to how electric engines replaced diesel engines in cars. The fundamental vehicle remains the same, but the power source transforms performance.
AI serves as the new engine for robotics, enabling machines to understand language, process visual information, and make decisions in complex environments. This integration creates opportunities for AI systems to interact with the physical world in ways previously impossible.
China’s Innovation Ecosystem
China’s AI landscape benefits from a unique combination of entrepreneurial energy and market dynamics. In cities like Hangzhou, approximately one in five people reportedly serves as a CEO, reflecting widespread enthusiasm for starting new ventures.
This startup mindset drives rapid experimentation. While most ventures may not survive five to ten years, the collective exploration helps identify promising directions and accelerate technological development. The competitive environment pushes companies to iterate quickly and maintain innovation momentum.
China’s market serves a dual purpose beyond traditional sales channels. It functions as a testing ground where new technologies mature rapidly through real-world deployment. This market-driven approach to technology validation helps identify which innovations can scale and which need refinement.
The Marathon Nature of AI Development
AI development resembles a marathon rather than a sprint. No single organization can maintain breakneck pace indefinitely, but collective progress through competition creates sustainable advancement.
Current advantages in AI development don’t create permanent barriers for new entrants. The field remains early enough that today’s leaders may not maintain their positions long-term. This dynamic environment creates opportunities for newcomers with fresh approaches and innovative applications.
The key challenge isn’t computing power—foundation models from leading companies already exceed human performance in many areas. Instead, creativity in application development represents the primary bottleneck. Most applications still follow the ChatGPT model rather than exploring novel use cases that leverage AI’s unique capabilities.
Building Sustainable Technology Businesses
Cloud computing’s success demonstrates how fundamental technologies create lasting business opportunities. Unlike typical business cycles that require reinvention every decade, computing infrastructure resembles electricity—a foundational service that remains relevant for generations.
This durability makes cloud computing particularly valuable. The technology supports whatever applications developers create, from e-commerce platforms to AI training systems. As AI becomes cloud computing’s largest customer, the infrastructure proves its adaptability to emerging needs.
The combination of data, computing power, and AI models creates new business possibilities that weren’t apparent when cloud computing first emerged. This convergence illustrates how foundational technologies often enable applications their creators never anticipated.
Practical Implications for Organizations
Organizations should focus on creative application development rather than competing on foundational model capabilities. Existing AI models provide sufficient capability for most use cases—the challenge lies in identifying problems worth solving and building applications that users actually want.
The rapid pace of AI development means short-term advantages rarely translate into long-term dominance. Companies should invest in understanding their specific use cases and building sustainable competitive advantages through application design and market understanding.
For new entrants, the current environment offers significant opportunities. While established players focus on existing successful approaches, newcomers can explore untested applications and serve underserved markets. The marathon nature of AI development means the race remains open to innovative participants.
The shift from toy problems to real-world impact represents AI’s coming of age. As embodied intelligence emerges and applications diversify, organizations that focus on creative problem-solving rather than technological competition will find the most sustainable success in this evolving landscape.