#thinking

Buffett, Munger, Duan Yongping, and Li Lu - mental models for decisions.

Systems Thinking วิธีคิดเชิงระบบที่คนทำงานและผู้บริหารต้องรู้ เพื่อแก้ปัญหาที่...

Understanding “media literacy” and its significance Media literacy is the ability to...

Should I take this job? Start this project? Invest in this tool? Say yes to this opportunity? We...

I spend most of my days inside a system I don’t fully trust. It tells me what to learn, when to...

Using an agent to build an inference engine from scratch, the project grew messier and performance degraded. The solution was always the same: stop, redefine the goal, open a new repo, and start over. The same model acted like a different entity. Microsoft simulated 200,000 conversations and found the same thing: a model on the wrong path won't turn back.

The best programmers aren't the ones who write the most code. They're the ones with the debugging mindset — relentless investigators who find root causes. AI can spot patterns, not systems.

Last month I said ignorance is an advantage. This month agents burned me four times: performance faked by orders of magnitude, lab machines bricked, optimization going in circles, and goals too ambitious to ship. Let me correct that: ignorance lets you try; common sense keeps you from getting scammed.

a follow-up to room to breathe and the recent heads-up about something longer, on how the space i refuse to fill with one more optimization is exactly where bigger, slower ideas finally get enough air to form

The new KMT chair visited the mainland, then came home and skipped traditional media — went straight to a livestream with an internet personality. She told Lei Jun her household runs entirely on Xiaomi. People like her remind me of the late Spring and Autumn period. The flip side of collapsing rites is a hundred schools contending. Same dynamic now.

why ai makes it feel like you can always do a little more, and why you still need room to breathe when you, not the model, are the one who has to review, understand, and ship the work

A few decades ago, AI meant machine learning. Three years ago, it was ChatGPT. This year, it's agents. The word never changed, but what it contains has been replaced three times over. Most people are still stuck in the previous generation's cognitive framework, discussing something that doesn't actually exist.