● AI Gold Rush
“Only the top 0.1% tech aristocrats survive” warning becomes reality: the K-shape turns into an H-shape, and the era of unipolar polarization arrives
The reason it feels off even when you just look at today’s news isn’t “a sense of deprivation”—it’s “the structure”
These days, you hear news like record-high stock prices, bonuses worth tens of billions, and real estate in Gangnam.
The emotion you feel isn’t just simple envy—it spreads into psychological deprivation like, “Why can’t I keep up?”
But what’s even more frightening is that this deprivation may be a sign that the actual acceleration of economic polarization in the global economy is underway, not just a matter of mood.
Key takeaway
As the speed of returns on capital (R) outpaces the speed of labor and growth (G), the structure that concentrates more wealth faster to the top is becoming entrenched.
1) The “K-shaped economy” worsens further and shifts toward an H-shape (or “the ladder gets cut”)
In the K-shaped economy, the line breaks, and the gap is confirmed not as a “feeling,” but as “statistics”
The K-shaped economy meant that there were groups that move upward while others fall downward at the same time.
But now it has intensified even more—so much so that the middle breaks off, and people even use the phrase “the ladder gets cut.”
What the actual numbers say
- Net worth gap expands up to 8 times
- Gini coefficient records the worst level since 2012
In other words, it’s not just a matter of “only some are doing well”—it allows for an interpretation that the structure of inequality itself is being solidified.
2) “Capital makes money quickly, while it’s hard for labor to keep up” (R > G)
R and G that Piketty talked about: what happens when the return on capital is higher than the growth rate
The original argument (Piketty):
If R (the required rate of return on capital) > G (economic growth rate), then people who have capital can build assets more quickly.
The key point is that this tendency repeats not as a one-off, but over long-term data.
The reality of assets that “increase by 1 million won a day”
What the example of the difference between sale prices and asking prices shows is that the gap in the rate at which effort (labor) and assets (capital) increase in price has become extreme.
Even if you work hard late nights and raise your salary, there are periods where the side that holds assets sees the price itself rise faster.
3) Why “tech aristocrats” are growing: big tech may not collapse—it could get bigger
Will big tech wobble in the AI era? → The conclusion of the article: “It’s possible it won’t”
The point is that it’s not like “if you use AI, you’re done”—it’s that those who have access to better models and infrastructure become more advantaged.
Why does being richer make you even more advantaged?
- Building or using high-performance models requires computing (electricity, GPUs, data centers)
- This accessibility then translates into gaps in cost/performance/speed
- As a result, it can lead to a “gap in intelligence → gap in productivity → gap in assets”
Before you make people smarter, this structure makes “people with authority” even stronger
Even if the number of users grows, from this perspective, the group that can afford training/inference costs and quickly secure optimized models can only be limited.
4) “Pressing-needle-shaped society (unipolar polarization)” scenario: an era where only the top 0.1% shoots upward
Overwhelmingly high growth rate: not 1%, but only 0.1% left in the lineup
What the graph interpretation suggests is a flow where an even smaller extreme minority among the top layer monopolizes the growth rate.
In such a society, the article describes it as dystopian and even uses the term “the era of unipolar polarization.”
To summarize
Instead of the middle class getting thicker, there are concerns that only the peak will remain thin.
5) There’s plenty of discussion about solutions, but the problem is “AI speed vs. institutional speed”
UBI, equal pay for equal work, robot tax, AI national dividends—ideas are all on the table
- Equal work, equal pay
- UBI (universal basic income)
- Raise funds through robot tax/AI tax
- Redistribute big tech profits (including in the form of national dividends)
But even if laws are created, “the time when they take effect” can be delayed
The core argument is this.
AI technology can be updated on a weekly basis, but institutional changes are incredibly slow—that’s the time lag.
Then, in the end, the structure can be strengthened where “the group that secures capital with AI first keeps being the winner.”
6) A warning that the past can repeat: the Industrial Revolution was not “labor liberation”
During the Industrial Revolution, expectations were “liberation from labor,” but the result was the opposite
Research is also mentioned where people expected labor to decrease in the era of textile machines and steam engines, but in reality work increased and wages did not improve significantly.
A repeating pattern in economic history
- The fruits of labor accumulate to a small number of capitalists
- Technological progress does not immediately spread to “everyone’s lives”
So the same kind of discussion can emerge in the AI era as well (a modernized twist of the “Engels effect”).
7) “Why AI isn’t free”: when marginal costs appear, incentives for monopoly grow
Software had low marginal costs, but intelligence is different
While past software had a structure where the cost of copying was close to zero, AI requires electricity, GPUs, and data centers for training and inference.
That means cost is involved in producing intelligence itself.
So who becomes the “parts” in the AI era?
Summarized as an analogy from the article:
In Capitalism 1.0, the “parts” were factory machinery; in the AI era, you can even consider a risky scenario where instead of “people who use AI,” “people replaced by AI” become the parts.
8) “So what should office workers do?”—two things presented in the original text + one additional (realistic summary)
Premise: people complain that work won’t decrease, but efficiency changes dramatically
When AI is introduced, it’s true that monitoring, inspection, and control tasks can increase—so the claim that “work doesn’t decrease” is also correct.
However, the key point in reality is that the ability to finish the same work in less time emerges.
Then the remaining time ultimately needs to be converted into “higher value work” or “product/productivity and personal asset accumulation,” and that’s the argument that follows.
Strategy 1) Get AI beneficiary stocks (infrastructure, bottlenecks) from a “stock price” perspective
The categories the article emphasizes especially are here.
- Power infrastructure
- Memory semiconductor chips
- Optical communications
- Data centers
The logic is that as the scale of AI investment grows, the “parts that become bottlenecks” have a higher chance of making more money.
Strategy 2) Before you’re exploited by AI, exploit (use) AI
The core is to “make the tool a habit.”
- Don’t stop at reducing simple tasks
- Use agents (delegated execution) to quickly produce actual outputs
- Reduce the 10 hours of work at your job to 1–2 hours, and use the leftover time for higher-level tasks in your main job and side job
A form that’s practical for office workers to implement right away (realize the original text)
- Structure current routine work (input–process–output)
- Define the problem first, then handle it quickly with “automation/agents”
- Use the leftover time to add things I create directly (websites/services/automated routines)
Strategy 3) Build your own assets (moat): what matters more than money is AI literacy
The conclusion of the article is strong.
Assets aren’t only money—they’re “the ability to use AI” itself
From this viewpoint, it can be the most realistic line of defense in an era of polarization.
Extracted separately from my own writing: “the most important content” (key summary)
1) The cause of polarization isn’t emotion (deprivation), but an R>G structure (capital returns outpacing the speed of growth and labor)
2) AI updates are too fast, while institutions are slow—so even if redistribution solutions come, the timing may be late
3) People who benefit in the AI era are those with “access to AI models and infrastructure” → ultimately there’s a stretch where bottleneck infrastructure becomes money
4) Office workers must “redesign their work” before they’re replaced by AI (use agents + produce outputs)
5) The final moat is likely to become an asset called “the ability to use AI” (literacy), not stocks/cash
Major message I want to convey
It’s hard to say that the AI era will definitely lead to a scenario where “everyone becomes better off.”
Instead, technology gaps → productivity gaps → asset gaps can sharpen top-heavy concentration even more.
So each individual doesn’t need to do anything overly grand.
Design a portfolio with an AI-beneficiary infrastructure viewpoint, and
internalize AI in your main job first to secure time efficiency, and
steadily build the moat (the asset) of AI utilization capability—that’s the realistic approach.
< Summary >
The global economy is moving into a more broken K-shaped form, worsening inequality, and the core cause is the structure where the return on capital (R) surpasses the growth rate (G).
In the AI era, a small group (tech aristocrats) with better model/infrastructure access becomes more advantaged, making unipolar polarization centered on the top 0.1% expand.
Solutions like a robot tax and UBI are discussed, but because institutional speed is slower than AI speed, the timing for redistribution may be late.
Office workers should look for opportunities from the perspective of AI-beneficiary infrastructure (power, memory, optical communications, data centers), secure efficiency through agent utilization in their main job, and need a strategy to accumulate “AI literacy” as personal assets.
[Related Posts…]
- AI Investment Strategy: How to Read the Next Cycle of Data Centers, Semiconductors, and Power
- Polarization Solutions in the AI Era: Why the Redistribution Debate Gets Delayed and Personal Survival Strategies
*Source: [ 월텍남 – 월스트리트 테크남 ]
– 상위 0.1% ‘기술 귀족’만 살아남는 시대, 당장 이 2가지를 시작하세요.
