Book: 20260115 to 20260503, "The Technology Trap" by Carl Benedikt Frey

20260115 - Preface

20260115 - Introduction

......adult male workers lost out: the share of children workers rapidly expanded, reaching about half of the workforce employed in textiles during the 1830s. p9
I never thought of this! Sending children to school killed two birds with one stone.

And more broadly, cascading competition among nation-states made it harder to align technological conservatism with the political status quo. The outside threat of political replacement became greater than the threat of rebelling workmen from below. p19
The political elites never back off unless they are under existential external threat. All the same.

Part I. The Great Stagnation

20260120 - 1. A Brief History of Preindustrial Progress - 33

What's the difference between labor-saving technology and enabling(capital-saving) technology?
Labor-saving technology use power from sources other than human. This power includes both energy and intelligence.

Fierce war and competitions crash the restrain from unions/guilds.
We need to thank China if Robotaxi and Optimus robots can become populated in US.

Will social welfare(including UBI) also help to spread labor-saving technical inventions? I think so.

20260202 - 2. Preindustrial Prosperity - 60

For example, the ecologist Jared Diamond has suggested that "forced to choose between limiting population or trying to increase food production, we chose the latter and ended up with starvation, warfare, and tyranny." p64
Jared Diamond got it wrong here. human always face this kind of choice in the past 100k years. The end of ice age and the accumulation of technogical progress made agriculture possible, then population started to grow significantly.

The rural industries that emerged were typically an off-season activity. many of workers living in the hinterland were both farmers and manufacturers. p70
Industrialization was birthed in the countryside.

20260203 - 3. Why Mechanization Failed

Incentives to mechanize hinge not on the freedom of the worker but on the price of labor, which remianed low in preindustrial societies. p75
The price of goods and services is decided by supply and demand.
The supply of goods and services is decided by social productivity, and social productivity is decided by the adoption of new technology.
The demand of goods and services is decided by population and median income(labor cost).
Machines are expensive. If median income is low, then the demand is low and it's not worth to use machines.

If US can use robots to auotomate all manufacturing, it's better to keep other countries relatively rich, so they are capable of purchasing goods and services from US, such as renting robots.
But US will not do it for free.
Only scarce resources are needed by US.

Geographic competition has gradually eliminated guild obstacles.
Geographic competition will do similar thing when AI arise.

The external threat of political replacement due to foreign invasion gradually became greater than the threat from below, as competition between nation-states intensified. p89
Nowadays, developed countries has no interest invading any other countries, thanks for the social welfare system. That means, most of the countries has almost no motivation to adopt labor saving technology.
Should we be grateful of the aggressiveness of Russia and China?

20260203 - Part II. The Great Divergence

But in the early days of industrialization, a great divergence, happened within Britain, too: wages stagnated, profits surged, and income inequality skyrocketed. p95
Can social welfare system solve this problem?

20260204 - 4. The Factory Arrives

Industrial Revolution also followed exponential growth. Just like the AI age?

Good to know the difference between charcoal smeltering and coke smeltering.

20260302 - 5. The Industrial Revolution and Its Discontents

Real wages started to grow after 1840, suggesting that there was an inflection point around that time. p135
When new technology getting matured, machines need more intelligent adults to operate and maintain. The critical factor I think it's still the relative fair political institution.
Many poor mirated to USA and Australia, that also helped.
More imported low price food also improved people's lives.

In the early days of industrialization, the gains from growth overwhelmingly went to owners of capital. p137

Attitudes toward technological change, as we shall see, are shaped by whether people can expect to benefit from it. p139

20260302 - Part III. The Great Leveling

Consequently, the rational response of labor was to allow mechanization to progress while minimizing the adjustment costs imposed on working people. p145
What can people do to minimize the adjustment costs durign AI age?

20260317 - 6. From Mass Production to Mass Flourishing

But while the twentieth century clearly saw the spread of some replacing technologies, most progress was of the enabling sort. p172
Maybe is more about the time it took to spead this technologies. If some technology can spread quickly, then even if it only improved the productivity for 10%, that still means 10% of people will get redundent.

20260317 - 7. The Return of the Machinery Question

Once the new became the familiar, attitudes shifted. p187
This makes sense.
The important one is the final attitude, not the inital attitude.

20260321 - 8. The Triumph of the Middle Class

It is therefore somewhat surprising that inequality (as measured by the overall Gini coefficient) plateaued around 0.5 and even fell slightly between 1870 and 1929, even though the top 1 percent share rose dramatically from 9.8 percent to 17.8 percent. p209

What will AI bring to us?

As we have seen, the race between technology and education does a good job of explaining trends in the labor market over the first three-quarters of the twentieth century. p217
No. It's about the relationship of supply and demand.
If there is under supply of Goods and Service, the inequality gap shrinks, or else it gets larger. The demand is partly decided by social welfare and international exportation.

Why are normal people in UK getting richer after 1870? The price of food dropped a lot, etc.

20260409 - Part IV. The Great Reversal, p223

Skill-biased technological change means that new technologies increase the demand for workers with more sophisticated skills, relative to those without such skills. p225

  1. Passive income doesn't follow this trend.
  2. AI will eliminate the demand for workers with more sophisticated skills.
    But we also need to consider "antifragile" problem, at least many years later.
    Pure productivity growth brings us fragility, just like globalization.

20260409 - 9. The Descent of the Middle Class, p227

Besides symbol-analytics services, Reich reckoned, there are also routine jobs and in-person services. p235
Thanks for Trickle-down effect, the high income symbolic analysts spend a lot of money to purchase the products from routine jobs, and the services from in-person services.
Thanks for Polanyi's paradox and Moravec's paradox, routine jobs and in-person services cannot be fully automated easily.
However, eventually, especially with the help from AI and Humanoid robots, less and less symbolic analysts are needed, and more routine jobs and in-person services are automated.
Capital assets will reap larger share from economical growth, and inequality gap will become larger.

20260410 - 10. Forging Ahead, Drifting Apart, p249

"the consequences of high neighbourhood joblessness are more devastating than those of high neighbourhood poverty......Many of today's problems in the inner-city ghettos - crime, family dissolution, welfare, flow levels of social organization, and so on - are fundamentally a consequence of the disappearance of work." p250
People need a place to store their attention. UBI/UHI needs to solve this problem! That's why drugs and screen are necessary to keep the society stable. Should I move to some area with less people? Or some place that less people may get drunk?

The Port Clinton story is regrettably a typically American one. p251
This also happened in Japan. It is going to happen in China.
But Europe seems got around it. Why?

While marriage has become les common across all spectra of society, the reason it has become so much more uncommon in Fishtown is that work has disappeared. p254
This is happening in China with the rise of Flexible Employment.

The flourishing cities of America have become nursery cities for innovation. But the rest is done abroad or by machines. p261
If innovation is driven by AI, what's going to happen? With the help of UBI/UHI, will people spread to countryside area?

Moretti estimates that each new tech job creates sufficient demand to support another five jobs in a given city. p262
Each manufacturing job lost, Moretti finds, costs another 1.6 jobs in the local service sector. p263
If the AI wave leads to mass unemployment among white-collar workers, the service sector will subsequently collapse.
Most manufacturing jobs have already been eliminated by automation; AI and robotics will only exacerbate the situation.
If white-collar, service, and blue-collar workers lose their jobs in succession, the arrival of UBI/UHI becomes inevitable.
The commercial real estate market, particularly the office sector, is also bound to crumble.

20260414 - 11. The Politics of Polarization, p264

Why didn't the working poor demand more redistribution? p271
Today, the main taxpayers are middle-class. They don't want to give more money to those who don't work.
And, elites are smarter. They always can figure out a way to keep their wealth.
But, if most of people lost their job, who would they do?

If people believe that they will eventually be made better off by technological progress, they are more likely to accept the churn. p287
It's critical to watch closely whether people have confidence in UBI/UHI, etc. in AI age.

20260428 - Part V. The Future, p296

20260502 - 12. Artificial Intelligence

However, for all their differences, these studies concur that unskilled jobs are most exposed to automation. p322
This book was written by 2019. Seven years later, today, 2026, it seems that no job is safe.
Medicines such as GLP-1 will reduce the demand of hospitals and surgeons, even less software engineers are needed.
When humanoid robots such as Optimus come out, that will be the end of jobs.

In 1987, when Robert Solow puzzled that "we can see the computer age everywhere but in the productivity statistics." p325
There was a few years of delay. AI expands much faster.
It's still at very early stage, but in a few years, most likely before 2030, AI will be good enough to handle most of the digital work.
60% of jobs in U.S. are white collar jobs.

Faced with falling wages and fading job options, some people may have chosen welfare instead of work, while others have struggled to find a job. p338
What level of welfare is acceptable to most of people? It needs to be enough to pay rent and daily expenses. That means it's equal or more than low income jobs.
This welfare needs to be UBI, which pay all citizens equal amount of money.

20260503 - 13. The Road to Riches

We have seen that most occupations that require a college degree remain hard to automate. p347
The author got confused with computer software and AI.
AI focuses on intelligence, while computer software focuses on automation.
Automation is the side product of AI. So no one can be immunized from the unemployment that caused by AI.

In this light, it stands to reason that as automation causes the incomes of many parents to vanish, it also diminishes the future prospects of their children. p351
What if the social structure is destined to be pyramidal? No matter how hard the government, society, and individuals strive, there will always be a vast number of people at the very bottom.
If it isn't a pyramid, what other possibilities are there? A spindle shape?
Nature abhors the middle class.
In the age of AI, the pyramid shape will be continuously reinforced until it becomes a thumbtack shape: the vast majority at the absolute bottom.
This returns society to the state humans experienced for thousands of years prior to the Industrial Revolution, aligning perfectly with the Pareto Distribution or the Matthew Effect.
The difference lies in the fact that, before the Industrial Revolution, those at the bottom were considered assets and resources. In the age of AI, they are purely consumers—neither assets nor resources.
But being a pure consumer is still meaningful, isn't it? Otherwise, if no one buys what is produced, the top billionaires won't make any money.
Wrong.
In an era where UBI (Universal Basic Income) and UHI (Universal Health Insurance) are ubiquitous, money becomes meaningless. The goal of the elite is not to "earn money," but to expand the resources in their hands.
In other words, they will reinvest all resources into the expansion of production.
With each passing year, the AI they wield will grow smarter; they will possess more and better (space-based) computing centers and robotics, and more (space-based) solar power stations.
Their wealth cannot be measured by "money."
Put simply: consumers provide no help to these elites.

It would be unwise to put too much faith in large-scale training efforts without proof of concept. p354
They didn't pay full attention in the training sessions, so they didn't gain much.
Attention is everything, to both human and AI.

Contrary to the anthropologist David Graeber's witty essay on "bullshit jobs," in which he claims that most people spend their working lives doing work they perceive to be meaningless, large-scale survey evidence shows the exact opposite. p357
This difference may be caused by psychological reason. As "Thinking, fast and slow" mentioned, people only remember the ending and the extreme points.

The bottom line is that regardless of what the future of technology holds, it is up to us to shape its economic and societal impact. p366
What's the best short-term and long term approach if AI takes over all jobs?

Comments

Popular posts from this blog

Book: 20250602 to 20250625, "Die With Zero" by Bill Perkins

Book: 20250703 to 20250815, "The Innovator's Dilemma" by Clayton M. Christensen

Book: 20250623 to 20250717, "What I Learned Losing a Million Dollars" by Jim Paul, Brendan Moynihan