You’ve probably noticed it too. The first two pages of Hacker News have become an endless loop of AI takes—this model beats that model, this founder predicts the end of software, that researcher warns about alignment. It’s exhausting. And it’s drowning out the real reason many of us come here: to learn how things work, not to hear someone’s hot take on what might happen.
I’ve been reading HN twice a day for nearly a decade. I’ve bookmarked over 50,000 links—a personal archive of everything that made me stop scrolling. And I’ve developed a system for what actually deserves to survive. My ranking is simple: if it’s visual, interactive, and explains a system from the ground up, it’s S-Tier. If it’s an opinion blog, it’s probably D-Tier or worse.
This isn’t just curation. It’s a statement about what kind of learning lasts. The best content doesn’t tell you something—it shows you how it works. An animated 3D walkthrough of a combustion engine? S-Tier. A 2,000-word essay on why Rust is the future? Usually C-Tier. The medium is the message.
So here’s my S-Tier list—the links that have survived a decade of my attention, sorted by the principle that seeing is understanding, and that understanding is the only thing that actually scales.
Want to really understand transformers? Watch them animate step by step. Want to get why k-means clustering works? Play with the data yourself. Want to trace a request from your browser to a server? Follow the 200ms journey.
There are 30+ links in my full collection, covering everything from elliptic curve cryptography to building a CPU from basic gates. Each one was chosen because it doesn’t just tell you—it makes you see. In a world of endless opinions, the only thing that scales is understanding.
And here’s the twist: I’m calling out the AI noise, but half of my S-Tier list is AI and ML explainers. The difference? They’re not opinions. They’re visual, interactive, and they let you build the intuition yourself. The problem isn’t AI content—it’s AI content that asks you to believe without showing you how. Curation is not about what you include; it’s about what you exclude.
So take this list. Use it. But more importantly, build your own system. The next time you see a post that promises to change your mind about AGI, ask yourself: does it show me something, or does it just tell me something? Your bookmark folder will thank you.
FAQ
Q: Isn't this just a list of links with a fancy ranking system?
A: It is a list, but the ranking system is the key insight. By prioritizing visual, interactive explainers, the author reveals a durable learning principle: you understand systems when you see them move, not when you read about them.
Q: What's the practical takeaway for someone who doesn't have time to read all these?
A: Start with the top three: the transformer explainer, the k-means visualizer, and the request journey. Each takes under 10 minutes and will give you a visceral understanding that no blog post can match. Then build your own curation habit.
Q: Doesn't the author's criticism of AI content undermine his own list, since half of the links are AI explainers?
A: Exactly. That's the point. The issue isn't AI content—it's AI content that is purely opinion-driven. The author's list proves that you can cover AI/ML topics in a way that actually teaches, not just hypes. The real enemy is the format, not the subject.