Posts on coffee_thoughts

Everything, Everywhere (in China) All at Once

Fri, 28 Aug 2026 09:53:16 +0800

A hundred flowers could bloom, the choice ones picked—but even that was no guarantee that a few of the most beautiful buds were to be trampled on later when tastes changed.¹

— Emily Feng, Let Only Red Flowers Bloom (2025)

I am currently reading Let Only Red Flowers Bloom, a banned book in China and apparently now in Hong Kong². The book depicts the government largely in a negative light—so it wasn’t surprising that it was banned—and shattered many of my prior assumptions about the CCP. As many reviewers have noted, it is one thing to look at abstract data and dry statistics, but encountering raw, heartbreaking personal stories taps into something far more primal and intimate. This is where the book becomes a much more effective medium than other sources (that I have read, for what it’s worth). Reading the intimate account of a human rights lawyer—from the height of their career to the harrowing, minute-by-minute reality of their detention and torture when they start to challenge the Party—leaves a mark that no summary of the 709 crackdown (en, zh) ever could.

AI, Globalisation, and the Black Death

Fri, 21 Aug 2026 09:22:41 +0800

I once conducted a simple thought experiment on epidemics and globalisation: if the COVID-19 pandemic had replaced the outbreak of the Bubonic Plague—otherwise known as the Black Death—in the 14th century, would it have become a simultaneous, planetary event of the kind the world experienced in 2020? Of course, the circumstances were vastly different in both advantageous and disadvantageous ways. People had much less knowledge of basic hygiene or medical practices, but populations were far smaller and more sparse. Population mobility was also significantly lower, as air travel did not exist and maritime transport was more modest. While long-distance movement existed at the time, such as overland caravans, Black Sea commerce, and Mediterranean shipping, it was slow and could not have crossed vast oceans. Thus, I believe it is not hard to imagine that the virus simply would not have spread far enough to have been a global epidemic¹. Everyday life was far more localised.

Why Does Indonesia Have Such Little Soft Power?

Mon, 10 Aug 2026 20:31:04 +0800

I recently saw a Reddit thread asking about Indonesia’s low cultural impact relative to its size. As the fourth most populous country in the world, it does seem a bit unusual for such a large country to have such little soft power.

I gave the question a good think, and below are my bullet points:

In general, low people mobility for many reasons: economic, cultural, and religious conservatism.
A middle class population that does not have enough capital to afford migration
- By comparison, India has a comparatively much larger upper class (via its caste-based hierarchy), which allows this very specific class of society to exert large Indian influence abroad.
  - Indian government also actively encourages the creation of ‘highly exportable’ individuals with its education system i.e. STEM education and heavy emphasis on English proficiency
- Indonesia’s diaspora relies mostly on hard labour exports e.g. construction workers, domestic helpers, which due to their socioeconomic status do not project much soft power overseas.
Generally conservative societies, due to a homogenous population and cultural & religious factors:
- Low incentives to study English which leads to low proficiency of the language, but English as lingua franca is a key point for globalisation
- Society which discourages external migration and puts high emphasis on collective good i.e. Gotong Royong
  - As opposed to a more pragmatic lndia/Thailand, which values upward mobility more than collective good (though it also exists).
- Although Indonesia is heterogenous in general, its regions are heavily localised i.e. people of similar backgrounds (race, religion, etc) are concentrated in the same few places.
  - Homogeneity disincentivises population migrations to elsewhere, as there would be less incentives to do so as well as pressure to retain the tight-knit communities they are living in.
Industry that focuses on raw material exports as opposed to processed goods (e.g. vehicles/consumer electronics) and/or non-tangible cultural exports (e.g. movies, food)
Protectionist economies create bad influencers
- Indonesia’s markets generally operate under Import Substitution Industrialization (ISI) model, which protects domestic industries from external competition by encouraging purchase of domestic goods
- Because of less competition, opportunity costs of exporting abroad becomes too high (e.g. to comply with strict foreign regulations), incentivising producers to produce domestically instead, as opposed to creating high-value specialty goods to be exported abroad.
- Because ISI is primarily enforced through heavy import taxes, populace will also be much more likely to buy locally produced goods, as they are much cheaper.
- As a consequence, Indonesia has little-to-no specialty goods exported abroad, which leads to less soft power:
  - Thailand has Monthong durians, specialty mangoes, and jasmine rice
  - Japan has its Crown melons, Shine Muscat grapes, Hokkaido dairy, etc
Government and populace that has failed to commoditise their culture (e.g. for tourism in Thailand)
- Thailand and Japan, for better and for worse, commoditises their culture heavily for tourism e.g. history, culture, religion, etc.
- An Indonesian outlier for this would of course be Bali, but similar successes of such scale have yet to be found in any other region of Indonesia.

If I had more time I would’ve reformatted this into a proper essay, but unfortunately I typed this on my phone early in the morning and by now I have lost motivation to turn it into a proper piece of writing, sorry!

Good Artists Copy Great Artists Steal

Tue, 28 Jul 2026 14:12:00 +0800

I have always been ashamed when I learn things; whether it is through reading or watching or whatever, I end up becoming ‘inspired’ by them. For me, it goes beyond simply drawing from broad concepts; more often than not, I find myself adopting the vast majority—if not the entirety—of the material itself.

The shame stems from my perception that it is wrong to steal someone’s work. But recently, as I have started documenting my thoughts, I realised that in the process of writing it down, I usually come up with my own ideas and opinions. I realised that simply processing it in my head normally allows very little room for me to fully digest the content and to interject it with my own thoughts based on my own knowledge. However, when you start taking your time to write it down, you realise that turns out you do have things to say about it, and you can synthesise knowledge from your working memory down into the piece you are currently writing about.

Diaries of a Polyglot

Sun, 26 Jul 2026 12:57:00 +0800

Ethnic identity is twin skin to linguistic identity—I am my language. Until I can take pride in my language, I cannot take pride in myself.

— Gloria Anzaldúa, Borderlands/La Frontera: The New Mestiza (1987)

I was born into a Chinese-Indonesian family. Being raised in a strictly Chinese household in Indonesia, my family primarily speaks two languages: Bahasa Indonesia, the national language, and Hokkien, a Chinese dialect originating from China’s Fujian Province—where my Chinese ancestors are from. Yet my background is far from unique; there are an estimated 2.8 million to 11 million Chinese-Indonesians in my country. Compared to the whole nation, this number comprises roughly 1.2 to 4% of the national population.. Though this percentage is relatively small, it also does not paint a complete picture, as we usually form tight-knit communities and are heavily concentrated only in several cities across the country. Most of this population therefore shares the same linguistic background as me, in particular the older generation, learning their ancestral language through their parents and their (mostly) Chinese neighbourhood, and learning the national language in school. The experience of navigating these two languages became a defining part of my daily life. In particular, I grew fascinated by the fluidity of code-switching—talking with my friends in Bahasa on the phone, only to immediately reply to my parents in Hokkien. Throughout my life, I would find out the extent to which I could push this ability, and how much I would end up benefiting from it.

A Skeptic's Opinion on Neural Scaling Laws

Mon, 20 Jul 2026 20:37:10 +0800

A plurality is not to be posited without necessity

— William of Ockham

I was recently reminded of the AI scaling problem, which is an argument against the premise that AI will continuously get better. Unless there is a major improvement in architecture, it is unlikely that AI will progress meaningfully (let alone revolutionarily) in the foreseeable future.

This might be a hot take, and certainly more of an educated guess rather than a theory-backed one, but perhaps it is time to rethink our approach in AI architecture. I am always reminded of Occam’s razor whenever I think about the prominent LLM architecture used today, i.e., the Transformer model. Of course, I lack the rigorous (and incredibly difficult) understanding of the most cutting-edge AI skillsets and knowledge. However, the entire architecture just seems overly convoluted with its massive scale. Not even leading AI researchers can concretely explain the precise efficacy of each transformer block. Although active research is being done on mechanistic interpretability, the landscape of this field is immature at best.

Reading Is a Hobby for the Privileged

Sat, 18 Jul 2026 20:55:00 +0800

What Orwell feared were those who would ban books. What Huxley feared was that there would be no reason to ban a book, for there would be no one who wanted to read one.

— Neil Postman, Amusing Ourselves to Death (1985)

I started reading very recently. In fact, it wasn’t until my sister gave me an e-reader around 2020 that I began seriously picking up books. As a child, I was never exposed to books, and I had always seen them as boring and intimidating. For some reason, I never had the motivation to read, and most importantly no one around me seemed to read either. It felt unnatural, to be seen picking up a book and reading it: a feeling that I think persists, if not more strongly, in the adult world I live in now. But recently I realised how remarkable it is to read: to be able to peer so easily into someone else’s imagined world through my own interpretation (hardly something you can do when you consume visual media). The words that had intimidated me for most of my life revealed themselves in all their elegance. Of course, countless studies have shown the risks of modern entertainment, and others have examined the literacy crisis and its consequences. They all raise important points, but I will leave them aside here for the sake of simplicity (go and read them yourself).

Pale Blue Dot Summary

Sat, 18 Jul 2026 11:25:00 +0800

This morning I read Pale Blue Dot by Carl Sagan. So far, I have only read the first few chapters, but I would still like to—given my extremely limited memory capacity—jot my thoughts down anyway.

The first few chapters establishes our insignificance, about how our existence is merely the product of a particularly lucky sequence of events, rather than the religious claims humanity has been preaching since time immemorial.

Nothing is easier than self-deceit.
For what every man wishes, that he also believes to be true.

My Notes on Speculative Growth and the AI “Bubble”

Fri, 17 Jul 2026 10:42:00 +0800

Original reading here.

My notes here are a massive oversimplification with a lot of content and details cut off. These are just what I, someone with no economics background, find interesting.

Assumptions

Productivity of AI

AI capital performs tasks previously done by labor. As it accumulates, productive capacity expands and income shifts toward capital owners, who have a stronger saving motive.

This only holds true if AI does actually expand ‘productive capacity,’ but then I suppose if there is no productivity increase then the entire mechanism, which is mainly economic, falls apart.