Book: 20260722 to 20260825, "Scale" by Geoffrey West

 

20260724 - The Big Picture

When the size is increased, the "cost" per unit and the "production" per unit are also changed.
The cost(consumption) is getting lower, and the production is getting higher.
Here, the production doesn't always mean good things. It could be something like crime rate.
There are a lot of factors affect those changes. When we remove a factor as bottleneck, another factor will become the bottleneck.
This is astounding. What is changing? New inventions such as solar panel, lithium batter, computer, EV, AI, robots, autonomous driving, smartphone, Internet, Starlink, etc.
A lot of limitations will disappear, and then many new bottlenecks will pop up.
No matter what, cities will become bigger and bigger. The living cost will be lower, and the average income will be higher.

Will this rapid growth of scale make cities more fragile? No.
Distributed energy(solar panel and lithium battery) and distributed intelligence(local AI) make each unit(person or family) more robust. Starlink, EV and tunnels make the communication system and the logistical system stronger.
We will hit the wall of growth someday, but not in the near future.

20260726 - The Measure of All Things

It's all about searching for the similarity. The dimensionless invariant numbers.
How about emergence? How is it affected by scaling?

20260802 - The simplicity, unity, and complexity off life

Their modern formulation is a general mathematical framework in which a quantity called the actions, which is loosely related to energy, is minimized. p115
The optimization itself is the mean, not the end. Life should be simple, but our aim should be high.

The reason that being healthy and robust equates with greater variance and larger fluctuations, and therefore a larger fractal dimension as in an EKG, is closely related to the resilience of such systems. p143
This also means that our lives need adventures.

20260807 - The fourth dimension of life

Less food intake can prolong our lifespan? If metabolism is the key, then surely GLP-1 can prolong people's longevity.
Many cases show that 72 hour fasting make use healthier, will there be some medicine do the same to us in the future?

20260812 - From the Anthropocene to the Urbanocene

Malthusian trap is real, but it only happens if fertility rate is high in very long period of time.
The reality is complicated.

1. There are too many babies in poor countries, and too less in developed countries.
2. Culture may change drastically in 10 years, which may cause fertility rate drop from 3.5 to 2(Japan, from 1950 to 1960).

My conclusion is that we don't need to worry about it.

20260812 - Prelude to the science of cities

If we don't know much about a city, then normally old district is better than new one, unless the new district is designed very well.

20260817 - Towards a science of cities

With each doubling of population size, a city needs only about 85 percent more gas stations - and not twice as many as might naively be expected - so there is a systematic savings of about 15 percent with each doubling. p272
The whole cost of the city, such as the number of gas stations, goes up at the slop of 0.85, and the opportunity(good or bad) goes up at the slop of 1.15.
When RoboTaxi and Humanoids get populated, unemployment rate is high, and people will spread out. Both slops(0.85 and 1.15) will get closer to 1.

In contrast, the car in Palo Alto was left untouched. p303
In big city, almost all neighbours are strangers. They may disappear next week, so it's not worth spending time talking to them.
That's the advantage of small town. In AI age, that's very important.

5, 15, 50, 150, 500, 1500. That's the limit of individual connections that our brain can handle.
In the age of AI, it's better to stay in a town with less than 1500 adults. Or else there are too many strangers, and it will be hard to trust each other.

20260822 - Consequences and predictions

The daily travel time we can tolerate is around one hour per day. And we don't want to visit faraway places often
Is this the reason that some cities have many skyscrapers?
Skyscrapers also caused traffic jam, which slow down traffic speed, and leads to shorter travel distance that people are willing to accept. Subway and train help a bit, but not that much.
How about self-driving? How about mass unemployment?

physical infrastructure(buildings, roads, etc.) vs socioeconomic dynamics (ideas, innovations, wealth creation, and social capital) p372
Living creatures need energy to do three things.

1. Keep healthy cells alive;
2. Create new cells to replace the dead ones;
3. Support the body to take actions;
   It's same to cities.
   A humanoid robot is good physician, lawyer, teacher, chef, gardener, handyman, accountant, musician, engineer, etc. Together with self-driving transportation, each person can get almost infinite energy and skill supply, although there will be mass or even total unemployment.
   So people will migrate from high dense population cities to rural area.

20260824 - Towards a science of companies

taking one year as the time slice, this says that the percentage of five-year-old companies that die before they reach six years old is the same as the percentage of fifty-year-old companies that die before they reach fifty-one. In other words: the risk of a company's dying does not depend on its age or size. p401
This is quite different from living creatures.

We found that the relative amount allocated to R&D systematically decreases as company size increases, suggesting that support for innovation does not keep up with bureaucratic and administrative expenses as companies expand. p409
Bureaucracy is the sign of death of companies.
AI: If you can choose only one public metric, go with the 5-year, industry- and size-adjusted excess SG&A residual; if you can choose three, add the gross profit per employee trend, along with the word frequencies of approvals, meetings, and departmental silos in employee reviews.

20260825 - The vision of a grand unified

Figure 78 in page 417 is amazing, with a minor error: each next cycle should be much shorter than the previous one, unless the X-axis is based on logarithm.

Kurzweil: "Within thirty years, we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended." This was written in 1993, so the prediction was that this will come to pass by 2023, which is no just seven years hence. I don't think so. p422
ChatGPT 3 was launched at the end of 2022. Not ASI yet.
But in 2027, it will be really close to ASI. And in many areas, it is ASI.
Astounding!

20260825 - Afterword

"The Fourth Paradigm: Data-Intensive Scientific Discovery": p443

1. empirical observation (pre-Galileo)
2. theory based on models and mathematics (post-Newtonian)
3. computation and simulation
4. or, 3.1, big data