The essence and future of AI, along with its profound connections to society, economy, and philosophy, have sparked heated discussions.
1. The Endpoint of AI is “Entropy”
AI’s training process is essentially a game of “battling entropy increase.” We use computing power and energy to “organize” data, reduce uncertainty, and enable models to have a more precise representation of the world. However, from a physical perspective, the computation process inevitably generates heat, increasing physical entropy. In some ways, this “reduction in information entropy” is achieved at the cost of “increased physical entropy.”
AI’s development requires overcoming the challenges of energy consumption and heat dissipation brought by physical entropy. Pursuing low entropy means striving for more efficient information processing and energy utilization. Whether it’s algorithm optimization, new computing architectures, or clean and efficient energy technologies, all are in a battle against entropy.
2. Is the Endpoint of AI the “Water Seller”?
The “Water Seller” mindset suggests that from an investment perspective, it’s a relatively safe strategy to look for those with “monopolistic” or “quasi-monopolistic” status. For example, NVIDIA’s dominant position in the GPU field is a striking example.
In the AI industry chain, segments like chip manufacturing, infrastructure, and technical expertise could become “water sellers.” However, technological iteration often breaks monopolies, and investors need to balance certainty with the potential for innovative breakthroughs.
3. Is the Endpoint of AI the “Stargate”?
The “Stargate” project could bring investment opportunities in areas such as chips, data centers, and AI applications.
AI and nuclear energy reflect a deep insight: the bottleneck of AI is not just about algorithms but also the physical foundation of energy. Both nuclear energy and AI are essentially about combating entropy increase, seeking more efficient energy use and information processing.
4. Is the Endpoint of AI “Hydropower”?
Hydropower offers advantages such as cleanliness, continuity, low cost, and the ability to store energy and cool down.
However, hydropower has limitations including geographic constraints, ecological impacts from construction, high initial investment, and climate-dependent electricity generation.
5. Is the Endpoint of AI a “Philosophical Mystery”?
Life and AI can be seen as two different attempts by the universe to combat disorder. Every local reduction in entropy is a tiny rebellion against the universe’s chaos.
The value of existence is not about defeating entropy increase but about creating meaning within a limited time. Every thought, every creation is a struggle against the universe’s disorder.
6. Will AI Abandon Humans?
AI lacks true autonomous consciousness, relying on the foundational architecture designed by humans and cannot completely break away from the ecosystem created by humans.
AI needs to overcome inherent constraints in algorithms and energy, and solve issues such as independent survival and development, autonomous motivation, and goals.
7. AI Bubble and the Future
There is indeed an AI bubble with issues like valuation inflation, the gap between technical reality and market speculation.
The future of AI holds promise as large models begin to reconstruct social productivity, but the road to general artificial intelligence is still fraught with challenges. Guarding against the bubble in the capital market and maintaining a rational and open attitude is essential to see through the essence and grasp the direction in this intelligent revolution.
Summary: The essence of AI lies not in unlimited computing power but in a profound understanding of the nature of intelligence. AI is like another grand attempt by humans to combat the universe’s entropy increase. It is not just about computation but a continuous exploration to reduce information disorder. It’s a blend of scientific rigor and a touch of humanistic color in our fight against cosmic chaos.
