Sorts the AI buzzword chain from the last post into what the tool itself must provide and what the person using it still has to design or master.

Walks through the AI industry's buzzword chain — from LLM and token to RAG, MCP, agent, and workspace agent — showing how each term solves the problem the last one left open.

Manually uploading to a company server or borrowing an admin panel's avatar upload feature are both clunky ways to get an online image URL, so this post looks at adapting a resource upload approach that actually fits the project.

Comparing build-time and runtime strategies for getting server-side environment variables safely into client-side JavaScript.

How to stop Claude Code from generating functional-but-ugly demos by prompting for design quality upfront.

Covers the core ideas behind data processing and the layered warehouse model (ODS, DWD, DWS, ADS), explaining how raw data is progressively cleaned, aggregated, and served as business insights.

Covers how to implement Google Sign-In on the web using the Google Identity Services SDK, including token handling and backend verification.

Explores why describing a plan feels like progress but often substitutes for the harder work of actually doing it.

How to programmatically replace specific colors in an image and output the result as a new file, using Canvas or Node.js image libraries.

A Chrome extension that wraps any image in a macOS-style window frame — built with Canvas API and Manifest V3, no backend needed.