Posts

Showing posts from February, 2026

What are Environmental Factors in Genomics?

Environmental factors in genomics refer to the conditions that shape how your genes are expressed, and they go far beyond physical aspects like weather, food, geography, or air quality. They also include social and economic influences, often described as privilege. Three key forms of privilege—financial resources, time, and food literacy—play a major role in determining whether someone can access healthy living conditions, rest, exercise, and informed dietary choices. The wealthier you are, the more likely you are to benefit from these privileges, while those in lower socioeconomic classes often lack them, making them more vulnerable to genetic predispositions for diseases such as obesity. Các yếu tố môi trường trong gen học là những điều kiện ảnh hưởng đến cách gen của bạn được biểu hiện, và chúng không chỉ giới hạn ở các khía cạnh vật lý như thời tiết, thực phẩm, địa lý hay chất lượng không khí. Chúng còn bao gồm các yếu tố xã hội và kinh tế, thường được gọi là đặc quyền. Ba dạng đặc...

Monogenic vs. Polygenic Approaches in Genomics

Modern genomics can test your predisposition to certain diseases or traits. Here’s how it works: scientists identify SNPs (single nucleotide polymorphisms), which are genetic markers found in less than 1% of the population. By sequencing your DNA, they compare your SNPs to those in existing databases and determine whether you may have certain genetic risks. However, this method has limitations. It is essentially a monogenic approach—matching one SNP in your DNA to one SNP in a reference set—while ignoring other factors that influence gene expression. Today, researchers increasingly use a polygenic approach, which considers multiple genetic variants together and incorporates environmental influences. This method acknowledges that your DNA sequence alone is not enough to predict health outcomes. To make accurate predictions, scientists must also include relevant information about your history, lifestyle, and environment. Gen học hiện đại có thể kiểm tra khả năng bạn mắc một số bệnh hoặc ...

Why Gene Ancestry Testing is Controversial?

There are thousands of genetic ancestry testing services online today, such as 23andMe. These companies claim they can identify where your ancestors came from and even tell you what percentage of your DNA is “European,” “Asian,” or other categories. However, these claims have been widely criticized. Your genetic sequence cannot directly reveal your exact ancestral origins. Instead, it can only show how similar your DNA is to reference samples in the company’s database. In other words, the percentages don’t represent how much of an ethnicity you “are.” They simply indicate how closely your genetic markers resemble those found in certain populations. Hiện nay có hàng ngàn dịch vụ xét nghiệm phả hệ gen trực tuyến, chẳng hạn như 23andMe. Các công ty này tuyên bố rằng họ có thể xác định tổ tiên của bạn đến từ đâu và thậm chí cho biết bạn có bao nhiêu phần trăm DNA “châu Âu”, “châu Á” hoặc các nhóm khác. Tuy nhiên, những tuyên bố này đã bị chỉ trích rộng rãi. Trình tự gen của bạn không thể t...

Being Born Again in a New Language

The reason I keep going with Vietnamese is not simply career prospects, friends, or culture. Of course, I love Vietnam and enjoy spending time with my Vietnamese friends, but I don’t think Vietnam is more special than any other country. I can imagine myself being just as fascinated by another language as I am with Vietnamese. The real reason is the feeling of being born again in Vietnamese. Since childhood, I was never fully the person I wanted to be—whether in Malaysia, Singapore, or Hong Kong. But each time I picked up a new language, I felt a subtle shift. My personality changes slightly in Chinese, English, and Malay. It wasn’t until I discovered Vietnamese—an entirely new language—that I felt like a blank sheet, ready to start over. The first time I was in Hanoi, everything felt fresh. No one knew me, no one understood me. That, to me, was freedom. I could redesign the kind of character, passion, and life story I wanted to show in this foreign land, through this foreign language. ...

The Self-Help Trap

Time is democratic. We all get the same 24 hours in a day. What differs is how we spend it—our pace, our efficiency, our priorities. That’s where the whole industry of time management comes in, neatly packaged under the label of self‑help. I’ll admit: I enjoy self‑help content on YouTube. It lifts my mood, sparks motivation, and gives me a comforting sense of hope and direction. But I also recognize the catch. Self‑help isn’t really about delivering the achievements it promises; it’s about selling the feeling that you’re already on the road to success. It’s less about the destination and more about the dopamine hit of believing you’re moving forward. So here’s how I’ve come to see it: self‑help works best as a motivation pill, not the actual work itself. Take it when you need momentum, when you need that push to get started. But don’t confuse it with the grind. Work is still work. No amount of motivational videos or clever productivity hacks can replace the hours you have to put in you...

What We Learn from Inventing Anna (2022)?

I just finished watching Inventing Anna (2022). It’s a great show, no doubt. The fact that the story is almost entirely based on true events is wild—and it also explains why there are no clear villains. Still, I’m convinced Anna is narcissistic, manipulative, and responsible for ruining many honest people’s lives. She is absolutely not innocent. One important takeaway for me is how the world today is fabricated by fame. You can literally fake fame into money and power, and Anna is the pinnacle of this game. She came from a grassroots Russian immigrant family in Germany, yet she styled herself brilliantly—fashionable, polished, and packaged perfectly for social media. She faked connections, pedigree, and relationships, using every trick to exchange them for a lavish lifestyle and to materialize her business out of nothing concrete. She also knew how to ride emerging ideologies and steer narratives in her favor. By repeating the story that she was a young female artist trying to break in...

Three Hacks for the Flashcard Grind

Discipline is undeniably important, but is that really all there is? I often notice that half the time I struggle to finish my Anki flashcards, while the other half I can effortlessly review 500 in one sitting. On good days, motivation carries me; on bad days, discipline keeps me going. Yet the equation doesn’t feel that simple. Are there small hacks that can push me to learn more? I’ve found three that help: first, start as early as possible—treat flashcards like a prerequisite before tackling the day’s main duties. Second, change the setting—cafés, libraries, even gyms work; you don’t have to stay in the comfort of your room or bed. Third, take advantage of the midnight reset—once the clock passes twelve and a new day begins, review then; there’s no need to wait until morning if you’re still awake. Kỷ luật chắc chắn là quan trọng, nhưng liệu đó đã phải là tất cả chưa? Tôi thường nhận ra rằng có những lúc tôi chật vật để hoàn thành các thẻ Anki, trong khi những lúc khác tôi có thể dễ ...

What's Up, Grinders!

I’m a big fan of a YouTuber called Evildea. He makes language learning videos, but not in the conventional way. He’s not a hyperpolyglot, not a linguist, and not selling courses. He’s just an enthusiast—a game developer who speaks English, Esperanto, Chinese, Spanish, and several minor constructed languages. He promotes three core principles for language learning: honesty, hard work, and fun. He also runs a community that investigates hyperpolyglots, exposing their nonsense—especially those who sell courses and make exaggerated claims. Evildea has a philosophy called the grind. Basically, you need to do your language drills every day before you can enjoy his videos. Language isn’t inherently hard; it’s just about grinding consistently over a long period of time. On top of that, he makes funny videos, does a bit of shitposting, and stirs up drama among other language YouTubers. But he also takes on serious challenges—for example, he’s been testing ALG methods using Dreaming Spanish for ...

What Gets Measured Gets Managed

One thing that makes me so obsessed with Anki is the stats. Not gonna lie—I have friends who think I actually work for Anki. Each morning, I open the app and grind for about an hour, then check how much time I spent and how many cards I reviewed that day. After that, I take a quick glance at the past week, month, and even my overall history, seeing my progress in a bigger picture. I’ll do a bit of math to figure out how much I’ve learned each day and each month on average, and at this pace, how far I can go by a certain point in time. I also track where I stand in this 10,000-card challenge. All these stats really paint a clear picture and motivate me to keep moving forward every day. In English, there’s a saying: “What gets measured gets managed.” Anki truly embodies this principle to the fullest. Một điều khiến tôi mê mẩn Anki chính là phần thống kê. Thật lòng mà nói, có bạn bè còn nghĩ tôi làm việc cho Anki. Mỗi sáng, tôi mở ứng dụng và học khoảng một tiếng, rồi kiểm tra xem hôm...

Why isn’t ALG the best for Vietnamese?

ALG stands for Automatic Language Growth. It’s a language acquisition method developed by Marvin Brown to teach Thai. The idea is that you don’t need to memorize vocabulary or analyze grammar—you don’t even speak at first. Instead, you just listen for hundreds of hours until words begin to flow naturally, and eventually you develop native-like fluency. I’m somewhat skeptical of this method for two reasons. First, you need a certain vocabulary base to even begin comprehensible input. Second, in practice, many languages simply don’t have enough materials that cater to learners at all levels, from beginner up to the point where you’re ready to speak. As a Vietnamese learner, I struggle to find input materials that match my level. Most of the content online is either too basic for beginners or fully native-level, with little in between. That’s why I think ALG may work for certain languages, but probably not for Vietnamese. ALG là viết tắt của Automatic Language Growth. Đây là một phương ph...

31 and What's Next?

I need more discipline and structure in my life because I’m getting one year older already, and I must be clear that I have three priorities: health, PhD, and language, in this order. In terms of health, I want to start two meals a day, walk 10,000 steps daily, and do simple weight training every day. In terms of my PhD, I need to spend at least six hours a day working on my thesis. In terms of language, I will review Vietnamese flashcards daily and take classes once or twice per week, start a commonplace book to jot down English words, and read and listen to more Malay to increase my input, which together may take two to four hours every day. My daily routine will be as follows: after midnight, I will review flashcards, then sleep before 2 am for six to eight hours; upon waking, I will continue flashcards if not yet completed, then prepare my first meal, shower, and start the day; I will devote six hours to research work, cook my second meal, do a light workout, walk 10,000 steps, and...

Why is Jollibee in Hong Kong the Best?

Today is my birthday, so I treated myself to my favorite fried chicken—Jollibee. Jollibee, without a doubt, serves the best fried chicken, and I think the Hong Kong branches are the best Jollibee outlets worldwide. That’s how highly I rate them. I’ve tried Jollibee in Malaysia and in its home country, the Philippines, but they still don’t compare to the Hong Kong ones. Why is Hong Kong Jollibee so good? I have my theory. First, Hong Kong is probably the city with the highest percentage of Filipino migrants, so the flavor should be closest to the original. Second, the Hong Kong outlets seem to have improved the recipe slightly, using better and richer ingredients since they can afford higher costs. Third, and most importantly, chicken in Hong Kong is sold separated into white meat (breast, ribs, wings) and dark meat (drumsticks and thighs), and all fast food restaurants here serve only dark meat—the juiciest parts of the chicken. That’s why Jollibee in Hong Kong tastes extra juicy, cris...

Like Any Other Day

Today is supposed to be a special day for me. I’m the birthday boy, turning 31. But honestly, I don’t feel much of anything. I’ll spend the day just like any other: practicing flashcards, sending emails, reading articles. I don’t expect to receive any birthday wishes either, since I don’t post about it on social media. And as we grow older, friends naturally have more important things to worry about, so birthdays often slip by unnoticed. That’s okay. It’s part of growing up. What matters is that I take better care of myself — something I sometimes forget. Hôm nay lẽ ra phải là một ngày đặc biệt đối với tôi. Tôi là nhân vật chính của ngày sinh nhật, bước sang tuổi 31. Nhưng thật lòng mà nói, tôi hầu như không cảm thấy gì cả. Tôi sẽ dành ngày này giống như mọi ngày khác: luyện flashcard, gửi email, đọc bài viết. Tôi cũng không mong nhận được lời chúc mừng sinh nhật, vì tôi không đăng lên mạng xã hội. Khi chúng ta trưởng thành, bạn bè tự nhiên có nhiều việc quan trọng hơn phải lo, nên sin...

To Grind or Not to Grind?

I am currently at 3,900 flashcards for my Vietnamese language. However, I feel quite stuck. I’ve been spending all my time practicing flashcards because I believe in two things. First, vocabulary is king—without it, you can’t communicate or understand. Second, sentence flashcards help me develop writing and speaking skills naturally, so I don’t need extra input practice. This makes things much easier: all I need to do is create flashcards, learn them, and review them. Then Vietnamese will come to me naturally. After 13.5 months, I can speak decent Vietnamese and handle daily communication quite well. However, when I watch clips of two Vietnamese speakers talking to each other, I sometimes understand almost nothing. That puts my listening at around a B1 level, while my reading, writing, and speaking are already closer to B2. This leaves me in a dilemma: should I continue building my vocabulary, or should I slow down and focus on listening practice? The former has given me great results,...

Why Valentine’s Isn’t the Best Idea?

Imagine you meet a girl you like months before Valentine’s Day. When the big day arrives, your feelings and connection haven’t matured enough to confess or make things obvious. You’re stuck in a dilemma: should you invite her out on Valentine’s or not? If you take the initiative too early, you risk killing the relationship. She might think you’re desperate or haven’t thought things through. But if you don’t ask her out, she might assume you’re not interested and give her attention to someone else. That’s the practical reason I’m not a fan of Valentine’s. Relationships need time to grow naturally, but Valentine’s imposes an artificial deadline that pressures people to rush or force things. I’d rather let connections develop organically. Of course, if you already have a girlfriend or wife, Valentine’s isn’t a problem. But for those still in the early stages, it can feel more like a trap than a celebration. Hãy tưởng tượng bạn gặp một cô gái mà bạn thích vài tháng trước ngày Valentine. Kh...

Elon’s Utopia or Elon’s Hell

Elon Musk envisions a future where machines and computer programs, powered by artificial general intelligence (AGI), achieve such extreme levels of automation that humans no longer need to work. In this imagined world, humanity could return to a Renaissance-like era, dedicating time to artistic pursuits and personal projects. Essentially, machines would take over all labor, leaving us free to enjoy life. I remain skeptical. Many global inequalities cannot be solved by automation—climate change, resource scarcity, geography, and more. Even more importantly, capitalism will not simply disappear. Today, capitalists extract profit from the surplus produced by human labor. But in a future where machines replace human workers, that surplus vanishes. Profit-making would then shift to other domains: data, energy, land, creativity, and beyond. This is why I doubt Elon’s utopia will materialize. Instead, new forms of value extraction will emerge, potentially in more exploitative ways. The means ...

Dietary Hierarchy

Have you heard of this? It used to be that wealthy people were more likely to be overweight, but today it is poorer communities that face higher rates of obesity. In the past, food resources were scarce, so the rich had access to more food, which often led to overnutrition. Today, however, food is abundant—sometimes excessively so. What is lacking is quality food and the time to maintain one’s health. Poorer people often rely on cheap, processed food and sacrifice time for work, while the wealthy can afford organic produce, whole foods, home-grown ingredients, gym memberships, and other health investments. This creates a kind of dietary hierarchy: there is “rich people’s food” and “ordinary people’s food.” To disguise this inequality, the emphasis is often placed on “green” or environmentally friendly consumption. While such food is genuinely better, behind it lies a deeper reality: health itself has become a form of capital, stratified along class lines. Bạn đã từng nghe về điều này c...

When do AI language models start to eat themselves?

I heard about a theory regarding the collapse of AI language models, often referred to as model collapse, and I find it fascinating. Have you ever played the game Catch Phrase? You line up five or more players. The first person receives a message and rephrases it before passing it to the next in line, and so on until it reaches the last person. Usually, the final message is nothing like the original. Some AI theorists worry that this same dynamic could cause language models to spiral into self‑referential loops. At the early stage—where we are now—AI is healthy because it is being trained on primary human data sources. However, the speed at which AI consumes data is far faster than humans can produce it. With AI assistance, more AI‑generated writing is also emerging, which is secondary data. In the foreseeable future, we may witness language models being trained increasingly on these secondary sources. This cycle could continue: tertiary sources, quaternary sources, and beyond—until th...

Go Get a Reference Manager

I had a library workshop today on setting up EndNote with my student. To be honest, I don’t usually use any reference manager, though I do take some notes in Zotero. I think we really do need to use a reference manager—it helps you organize all your citations and generate a bibliography in any format with just one click. It’s much quicker than asking ChatGPT to edit for you. A reference manager can also serve as a tool within a broader knowledge management system. I can use both Obsidian and Zotero at the same time to track all the readings and notes I’ve done. I need to build that system as soon as possible—in fact, I should start tonight. Hôm nay tôi có một buổi workshop tại thư viện về cách thiết lập EndNote cùng với sinh viên của mình. Thật lòng mà nói, tôi thường không dùng phần mềm quản lý tài liệu tham khảo, mặc dù tôi có ghi chú trên Zotero. Tôi nghĩ chúng ta thực sự cần sử dụng một phần mềm quản lý tài liệu tham khảo—nó giúp bạn sắp xếp tất cả trích dẫn và tạo danh mục tài liệ...

Why don’t some Vietnamese words match their sounds?

Vietnamese, I argue, has an oral origin. Historically, it adopted the Chữ Hán script, the Chữ Nôm script, and later the Chữ Quốc Ngữ script to represent its sounds. Yet this adoption was never perfect. There are moments when these foreign script systems could not fully capture the corresponding Vietnamese sounds. Out of this tension, certain “deviants” emerged. Ironically, these “deviants” became crucial in preserving the accuracy of Vietnamese original sounds, ensuring that the language’s oral essence remained intact even as scripts struggled to contain it. Tôi cho rằng tiếng Việt có nguồn gốc truyền khẩu. Trong lịch sử, nó đã tiếp nhận chữ Hán, chữ Nôm, và sau đó chữ Quốc Ngữ để ghi lại âm thanh của mình. Tuy nhiên, sự tiếp nhận này chưa bao giờ hoàn hảo. Có những lúc các hệ thống chữ viết ngoại lai này không thể phản ánh đầy đủ các âm thanh tương ứng của tiếng Việt. Từ sự căng thẳng này, một số “biến thể” đã xuất hiện. Trớ trêu thay, chính những “biến thể” này lại trở nên quan trọng...

Data Darwinism

I recently came across the idea of data Darwinism. Let’s unpack it. Darwinism is the hypothesis that only the fittest survive. Organisms evolve in ways that maximize their chances of survival, which explains why they are what they are today. Data Darwinism emerges from the age of AI. AI consumes vast amounts of data every day, but it must select the most representative data to present to us. Over time, peripheral or less common data gets pushed aside as AI increasingly relies on secondary, filtered layers of information. The danger, according to this hypothesis, is that human beings may eventually stop hearing scattered or minority voices, because machine learning systems decide to screen them out. That prospect is troubling. Tôi gần đây bắt gặp một ý tưởng gọi là data Darwinism. Hãy cùng phân tích. Darwinism là giả thuyết cho rằng chỉ những kẻ thích nghi tốt nhất mới có thể tồn tại. Các sinh vật tiến hóa theo cách tối đa hóa khả năng sinh tồn, và đó là lý do chúng có hình dạng như ngà...

Prompt Engineers or Machine Yappers?

I’ve got a bone to pick with prompt engineering. Let’s be honest. It’s basically writing clear instructions for an AI chatbot. That’s not engineering, that’s glorified machine yapping. No coding, no machine learning, no AI expertise, yet suddenly everyone slaps “engineer” on their title like it’s a badge of honor. By that logic, since I chat with friends every day, I should be a verbal engineering specialist. See how absurd that sounds? The inflation of titles is getting out of hand. So, dear prompt engineers, maybe drop the pretense and embrace honesty. You’re not engineers. You’re professional machine yappers. And honestly, that’s a much better fit. Tôi có một vấn đề với prompt engineering. Thật lòng mà nói, nó chỉ đơn giản là viết hướng dẫn rõ ràng cho chatbot AI. Đó không phải kỹ thuật, mà chỉ là machine yapping được thổi phồng. Không có lập trình, không có học máy, không có kiến thức AI, vậy mà ai cũng tự gắn chữ “engineer” lên danh xưng như một huy hiệu danh dự. Nếu thế thì, vì t...

What We Learn from Apple Cider Vinegar (2025)?

Apple Cider Vinegar (2025) is a very relevant TV series today. By “relevant,” I mean it reflects the problems of a post‑truth, neoliberal society gone wrong. Belle is a fraud who fakes cancer and uses it as personal branding to run a startup that recommends curated meal plans for cancer patients. She also commits charity fraud and promotes alternative medicine that literally costs lives. The story follows how Belle is eventually taken down, and the storytelling is very strong. I want to highlight three things the series brings up. First, entrepreneurial and startup culture today nurtures egoistical, narcissistic individuals who use “humanity” as a brand to sell. Second, the privileged eco‑green alternative lifestyle—whole, organic, gluten‑free, vegan decaf soy lattes with almond milk—functions as a way to greenwash privilege and create a dietary hierarchy in society. Third, as Byung‑Chul Han has described, we live in a compulsory positive society: when it comes to cancer, Milla and Luc...

Humanities Research in the Age of AI

I’ve been reflecting on the role and future of humanities scholars in the age of AI. In two or three years, I’ll be entering the job market myself, and the pace at which AI is advancing makes me wonder whether many of us risk becoming irrelevant. After all, AI can already read, write, analyze, and even teach. It has access to tools and knowledge across disciplines. Although the research it produces is still far from meeting academic standards, it’s not hard to imagine that in the near future it could replace much of what we do. So what should we do? One mistake I think we’ve been making is gatekeeping the use of AI in research writing. We remain purists, insisting that scholarship should be as close to 100% human-authored as possible. Of course, writing by hand—or “thinking in ink”—is a vital part of learning, but clinging to that alone doesn’t address the deeper issue of AI encroaching on our intellectual territory. Instead, I’ve been sketching out a different model. Rather than frami...