brandonwie.dev — Software engineering insights, tutorials, and learnings

brandon@seoul:~ $ whoami --verbose

Software engineer writing about backend systems, DevOps, and lessons learned building real products.

Full-stack AI engineer building production systems — film school grad turned engineer, now building AI products at Playtag. This blog collects what I've learned along the way.

Projects →All postsAbout Brandon

brandon@seoul:~ $ ls -l ~/work

brandon@seoul:~ $ ls -lt ~/posts | head -10

Recent Posts · 10/168

#datecattitle / description

01Oct 29, 2026devops

Anthropic Prompt Cache TTL + Cost Mechanics

Anthropic silently dropped Claude Code's prompt-cache TTL from 1 hour to 5 minutes around early March 2026. Without explicit awareness, idle gaps ≥5 min between messages evaporate the cache and force a full cold cache-write on the next message — pricing it at 1.25× base input on the entire conversation prefix.02Oct 29, 2026general

Claude Code Multi-Profile HUD Setup

Running Claude Code with multiple accounts requires careful HUD configuration to show correct per-account usage stats. Here's how to fix cross-profile data leaks.03Oct 22, 2026aws

AWS VPC Networking Fundamentals

Comprehensive guide to AWS VPC networking: CIDR notation, subnet design, NAT Gateway placement, route tables, and complete Terraform examples.04Oct 22, 2026devops

A Harness That Fixes Itself - And Prunes Its Own Fixes

The most expensive agent mistakes are not the dramatic ones.05Oct 22, 2026devops

Claude Code: Shared + Personal AI Config Pattern

Split AI instructions into committed (shared) and gitignored (personal) layers06Oct 22, 2026devops

The pre-commit hook race that put my files in someone else's commit

Two sessions committing to one repo, a slow pre-commit hook, and `fatal: cannot lock ref HEAD`. The loud failure is the easy one — the quiet failure hands your staged files to the other session's commit under its message.07Oct 20, 2026ai-ml

Claude Code Agent Teams

Experimental feature for orchestrating multiple Claude Code instances as a coordinated team with shared task lists and inter-agent messaging08Oct 20, 2026devops

Wrap Skill Follow-Up Persistence Architecture

When a session-state dashboard regenerates from a single source (today's journal), unresolved follow-ups from prior sessions vanish silently on every rebuild. Compounded with single-source discovery and conversation-only mentions, follow-ups disappear three ways at once. The fix is a 4-layer architecture.09Oct 6, 2026frontend

`_emphasis_` does not close against a Korean particle

CommonMark will not close an underscore emphasis run when Hangul sits on both sides of the closing delimiter, so the underscores stay literal. Asterisks have no intraword restriction and still close.10Sep 20, 2026ai-ml

AI Code Review Confusion Patterns

Thirteen distinct ways Claude, Copilot, and Codex behave on PRs — ten failure modes plus two productive behaviors to amplify, plus an analyst-side error class. With detection signals and the empirical tiebreaker that resolves factual disagreements.

... +158See all 168 posts →

fzf --preview · 01

devops · Oct 29, 2026

Anthropic Prompt Cache TTL + Cost Mechanics

Anthropic silently dropped Claude Code's prompt-cache TTL from 1 hour to 5 minutes around early March 2026. Without explicit awareness, idle gaps ≥5 min between messages evaporate the cache and force a full cold cache-write on the next message — pricing it at 1.25× base input on the entire conversation prefix.