Summary of Key Points
This article focuses on two major bottlenecks in the development of the AI industry: first, the scarcity of time (while AI increases the output of knowledge workers, people's time is limited, and the additional output cannot be easily absorbed); second, insufficient purchasing power (AI replacing mental labor impacts the income of the middle class, leading to reduced consumer spending and creating a vicious cycle of "increased productivity → unemployment → decreased consumption"). The article provides solutions to these issues by citing historical examples and economic principles (such as Jevons' Paradox and growth models), and suggests that the realization of AI's benefits will go through two phases: a short-term period of discomfort (supply exceeding demand) followed by a long-term boom (demand and systems adapting to each other). It ultimately concludes that there is no need to worry about demand and purchasing power issues in the long run.
The Two Major Hurdles for AI: Can't Sell It, or Is It Too Expensive?
The biggest concerns regarding AI development are not its technical capabilities but rather two practical barriers to widespread adoption:
- Time Bottleneck: For example, if a knowledge worker's output doubles with AI, who will buy the additional services? You can only watch videos for up to 1.5 hours a day; even if the AI-generated content is engaging, people won't watch for 10 hours. Companies would need to see a tenfold increase in sales to justify investing in such services—otherwise, the excess computing power will be wasted.
- Purchasing Power Bottleneck: Previous technological revolutions replaced physical labor (e.g., machines replacing workers), but AI affects mental labor (white-collar jobs, programmers, designers, etc.), directly impacting the income of the middle class. If many people lose their jobs or see reduced salaries, even if AI makes services more affordable, they may not be able to afford them, leading to a cycle of "increased productivity → decreased consumption."
Jevons' Paradox: The More Efficient the Technology, the More Demand?
The first bottleneck is not insurmountable. Similar issues have emerged throughout history, and the solution lies in the fact that technological progress creates new demands—this is what Jevons' Paradox describes:
- For instance, Ford's assembly line greatly increased automobile production. Initially, there were concerns about where to sell all the cars, but the lower cost made them more accessible (a car could be bought with three months' salary), turning them from a luxury to a necessity and sparking new industries like highways, gas stations, and suburban housing, thus increasing demand.
- In the internet era, people feared an overload of information, but new forms of consumption emerged (short videos, knowledge-based services, live shopping). Those without money but time consume short videos, while those with money but no time pay for knowledge-based content. Demand has not decreased; it has become more diverse.
Overcoming the Purchasing Power Barrier: A Combination of Technology and Credit
The core issue is the gap between rising productivity and purchasing power. The solution lies in the nature of economic growth: technological progress + expansion of credit.
The article uses a fairy tale to illustrate this point: Two characters, Bear and Pig, had only one dollar and could not make transactions. A scientist (technology) doubled bread production, and a banker (credit) provided loans for Pig to expand production and buy more bread for Bear. Although they had to share the profits with the banker and the scientist, the overall economic size increased, benefiting everyone.
- Technological Progress: AI increases total supply, directly boosting GDP.
- Credit Expansion: Banks offer loans for businesses to purchase AI tools and consumers to access AI services in advance, enabling them to afford these services.
- Debt Ratio: Technological progress expands the denominator of GDP, reducing the debt ratio (for example, if GDP doubles from 100 to 200 and debt remains at 100, the debt ratio drops from 100% to 50%).
The Release of AI Benefits: Pain First, Then Boom
The realization of AI's benefits will not happen overnight. It will go through two phases:
- Short-Term Discomfort: Supply grows faster than demand, leading to unemployment and income inequality (e.g., some highly paid professionals spend a third of their time on inefficient tasks like making PPTs, which AI can automate).
- Long-Term Boom: As demand and credit expand, systems adapt (e.g., shorter working hours, improved social security, new wealth distribution mechanisms), and the full potential of AI is realized.
For example, Accenture mentions that tokens are currently used for inefficient tasks like converting PDFs to PPTs. In the future, AI will transform organizational processes, eliminating manual tasks and freeing up resources for more valuable work.
Investment Implications for AI
The article offers the following advice on investment:
- Long-Term Outlook: There is no need to worry about demand and purchasing power. Human desires are unlimited; the only limit is supply. As a technological revolution, AI will create new demands and increase total supply, just like the internet did.
- Short-Term Focus: The realization of AI's benefits takes time (the steam engine took 100 years, the internet took 10 years). Investors should focus on the speed at which demand is adopted (e.g., how quickly AI will transform organizational processes).
In summary, AI is not a fleeting trend but, like previous technological revolutions, will undergo short-term adjustments before fundamentally changing society's production and consumption patterns.