Hello! I'm your financial analyst friend. Today, we're going to discuss an article that contains a lot of information. It brings together the perspectives of two seemingly unrelated but actually closely connected "big players": the economic model of Anthropic, a top AI company, and the social warnings from Bill Gates, a technology giant.
To make it easier for you to understand without having to wade through complex economic models, I've broken down this long article into five key points. Let's first see what the article is about and then delve into each of them.
📝 Summary of the Key Points: The Reality of "Divisions" in the AI Era
The main argument of this article is quite stark: the prosperity brought about by AI may not automatically translate into increased income for ordinary people.
- Anthropic says: If AI becomes widely adopted, the GDP of the United States could soar by 32%, leading to high economic growth. However, at the same time, the unemployment rate could soar to 12%, and the wages of ordinary workers might not only stagnate but also decrease by 11%.
- Bill Gates says: This situation where the country gets richer while individuals get poorer is dangerous. Our current systems (taxation, education, social security) are designed for humans to work for. Now that machines are doing the work, these systems are not adapted, which could lead to social divisions.
- Conclusion: The biggest challenge in the future is not how to create smarter AI, but how to distribute the money generated by AI and how people can find dignity and value when work is no longer the only source of income.
---
🔍 In-Depth Explanation: Five Easy-to-Understand Dimensions
1. **The Decoupling of Growth and Employment:** The Pie Is Bigger, but the Knife for Cutting It Has Changed
【Easy-to-Understand Explanation:**
We used to think that when the economy improved, everyone would get a job. For example, if a factory expanded production, it would hire more workers. This is what we call "growth driving employment."
But AI has broken this rule.
- In the past: Companies wanted to make more money → They bought equipment and hired employees → Employees got paid → Employees spent money → Companies made more money. It was a closed loop.
- In the AI era: Companies want to make more money → They buy chips, computing power, and use AI to replace employees → Profits increase, but the number of employees decreases → Total wages don't increase much → People have less money to spend → Companies can't sell as many products.
Key Point: Anthropic's model shows that in extreme cases, although GDP increases by a third, labor income (the total wages of everyone) hardly increases, and even the wages for cognitive jobs (like coding and analysis) may decrease by 31%.
What does this mean? The country is financially wealthy, and the stock market is high, but ordinary people's pockets are not getting richer; in fact, they might be getting poorer. This is what's called "phantom GDP" – the numbers look good, but ordinary people don't feel the benefits.
2. **The Breakdown of the Career Ladder:** Young People Are Losing Their Jobs Before They Even Start
【Easy-to-Understand Explanation: Many people worry that AI will take away jobs from senior experts, but the most dangerous impact is on junior positions.
- Current Situation: Lawyers, programmers, and analysts all start with basic tasks like researching, writing code, and making simple reports.
- Impact of AI: AI is excellent at doing these standardized, repetitive tasks. To save money, companies use AI to replace these jobs.
- Consequences: Only a few senior experts remain in the company, and the middle and lower levels of the workforce disappear.
- For young people: You graduate, but there are no entry-level jobs for you to gain experience. You can't start with basic tasks because AI can do them faster and cheaper than you.
- For companies: They save money in the short term, but what will happen in five years when the senior experts retire? Where will the new generation of experts come from if no one has the basic skills?
Key Point: This is not just about unemployment; it's a breakdown in the talent training system. Education levels are rising, but entry-level jobs are disappearing, leaving young people with knowledge but no opportunities to gain experience, leading to a new type of "structural unemployment."
3. **Uneven Wealth Distribution:** Those Who Own AI Own the Future
【Easy-to-Understand Explanation: If AI can do the work, where does the money go?
- Traditional Model: Money goes to the people who do the work (wages).
- AI Model: Money goes to those who own the machines (capital returns).
Anthropic's data is alarming: Capital income (what shareholders and investors earn) is 81% higher than in the baseline scenario, while labor income has almost stopped growing.
This leads to a new class divide:
- Those with AI assets: People who own NVIDIA stocks, data centers, or develop AI models. Their wealth will grow like a snowball.
- Those who rent AI assets: Most workers. Even if you learn to use AI, you're essentially paying rent (subscription fees, token fees) to the AI platforms. You're not sharing in the huge profits from the underlying technology.
Key Point: The future wealth gap will not be just between high-skilled and low-skilled workers, but between those who own the means of production (AI, computing power, data) and those who sell their labor. If nothing is done, wealth will be highly concentrated in the hands of a few tech giants and capital owners.
4. **Policy Cannot Rely Solely on "Retraining": Don't Make Systemic Problems Personal Responsibilities**
【Easy-to-Understand Explanation: The government often says, "Don't panic, learn new skills, and get retrained."
But this article argues that retraining is a band-aid, not a solution to the problem.
- Logical flaw: If 100 people used to do a job, now 10 people plus AI can do it. Even if you retrain all 90 people to use AI, companies still only need 10 experts.
- Real issue: The problem is not that you can't use AI; it's that the overall demand for labor has decreased.
What to do? The article proposes three approaches:**
1. Short-term: Retraining and job switching (necessary but not sufficient).
2. Medium-term: Shortening working hours. If machines are more efficient, why not work fewer hours and enjoy more leisure time? Convert productivity gains into free time instead of layoffs.
3. Long-term (most important): Universal capital participation.**
- Bill Gates suggests taxing AI (e.g., taxing tokens) to fund subsidies.
- The author believes that a more fundamental solution is to give everyone a share in the benefits of AI. For example, using pensions and public funds to invest in AI infrastructure or implementing employee stock ownership so that ordinary workers can share in the company's increased profits.
- Core logic: Shift from a "relief logic" (giving money to survive) to an "ownership logic" (letting people become shareholders and share in the profits).
5. **Human Areas of Retention:** Some Jobs Are About Dignity, Not Just Efficiency
【Easy-to-Understand Explanation: This is a poignant point raised by Bill Gates.
- Misconception: We often think that if machines do things better and cheaper than humans, humans should step aside.
- Reality: In some jobs, human involvement is essential for value.
- For example, when a doctor tells a patient about a terminal illness, a cold AI screen might convey the same information, but the patient will feel indifferent.
- In elderly care, robots can feed and turn patients over, but what the elderly need is human interaction, emotional support, and a sense of dignity.
Key Point: Society can choose to retain human dominance in certain fields (medicine, education, justice, psychology). This is not anti-technology; it's about redefining the boundaries of efficiency.
- New model: A hybrid model where humans are ultimately in charge: AI handles data and provides advice (assisted decision-making), but the final decisions, responsibilities, and emotional connections must be made by humans.
- New value of humans: Human value comes from taking responsibility, building trust, and making value judgments, not just from repetitive tasks.
---
💡 Takeaways for Ordinary People
1. Don't Just Focus on Learning AI Tools: Learning to use tools like ChatGPT to write code is important, but if everyone does it, your competitiveness doesn't increase. Instead, think about: How do you define problems? How do you take responsibility for the results? How do you organize resources? These are high-level skills that AI can't replace yet.
2. Pay Attention to Asset Allocation: As labor income may decrease and capital income may increase, the importance of owning assets (stocks, funds, property, etc.) is rising. Consider legal financial tools to share in the benefits of technological development, not just rely on a fixed salary.
3. Redefine the Meaning of Work: If future jobs become unstable or working hours decrease, think in advance about what else you can rely on for social recognition and self-worth. Community involvement, creative activities, and interpersonal relationships may become as important as wages.
4. Be Alert to "Losses in Prosperity: Even if the country's GDP grows, if you don't feel the benefits, don't ignore it; this is a structural issue. Stay informed about public policies because future tax and social security reforms will directly affect your quality of life.
In a nutshell:
In the AI era, the greatest danger is not poverty but being "abandoned by prosperity." Our goal should be not just to learn how to use AI but to drive systemic changes to ensure we are both users and beneficiaries of its benefits.