Notes
Knowledge notes, build logs, and things worth writing down.
AI workflow ဆိုတာ မနက်အိပ်ရာနိုးတာနဲ့တူတယ်။ trigger, data, AI step, human check, output — ဒီငါးဆင့်ကိုနားလည်ရင် AI automation တစ်ခုလုံးကိုနားလည်ပြီ။ beginner-friendly storytelling ပုံစံနဲ့ ရှင်းထားပါတယ်။
Read note →Content collections turn markdown files into typed, queryable data. Here's why they're the right architecture for a blog or notes section, and how to set them up from scratch.
Read note →Getting dark/light themes right is harder than swapping background and foreground colors. Here's what breaks, what the FOUC problem is, and how the token system in this portfolio handles it.
Read note →CSS animations are fine for hover states. For anything that needs to feel alive — sequenced, staggered, timeline-driven — GSAP wins every time.
Read note →The grid background, monospace everything, dashed borders, glowing dots — all of it is pure CSS. Here's how the visual language of this portfolio works.
Read note →Most AI memory implementations fail because they treat memory as retrieval. Real memory is about what gets recorded, what gets forgotten, and what gets surfaced at the right time.
Read note →Most prompt advice is optimized for demos. Production prompts face adversarial inputs, model drift, edge cases, and context limits. Here's what actually works.
Read note →Most side projects die because they try to become products. The ones that survive stay small, solve one thing precisely, and resist the urge to grow.
Read note →JSON configs break silently. YAML configs have footguns. TypeScript configs give you autocomplete, validation, and refactoring — for free, with no extra tooling.
Read note →The Next.js default is to reach for SSR. The right answer is usually SSG. Here's a framework for deciding which rendering strategy to use for each page.
Read note →Terminals have survived for 50 years because they get a few things exactly right. These principles aren't just for CLI tools — they make any interface better.
Read note →