posts/ — archive_□×
Getting the First Agent Running in Paperclip
Second-Resolution Timestamps Miss Concurrent Edits
Security Headers Missed the HTML Pages
My Search Title Boost Was Applied to the Wrong Columns
My Logger Redacted `monkey` Because It Contains `key`
My Tests Skipped Concurrency and Authentication Wiring
My Quality Tool Gave Missing Checks a Perfect Score
The Billing Timezone Didn't Need a Dropdown
Unequal Search Weights Can Still Tie
When a Dependency Check Can't Answer
My Pull Request Was Green. My Tests Never Ran.
Pinned Fixtures Need a Separate Freshness Check
An Unsafe Autofix Can Change What a Test Covers
The Published Import Wasn't the One in the Source
An Empty Environment Variable Still Needs Validation
The Test Checked the Arguments, Not the CLI Contract
Install-Script Permissions Can Outlive a Dependency
Four Environment Settings Never Reached the Logger
The Rate Limiter Trusted a Client-Controlled Header
The Tests Made a Case for Deleting the Shell Script
The Upload Allowlist Didn't Bind the Content-Type Header
Removing a Dependency Has More Than One Meaning
Bun Was Faster. I Still Kept the AWS SDK.
An SVG Is Not an Image
Two SQLite Libraries, One Deleted Password
Your PDF Remembers What You Deleted
Looping Is Not in the GIF Spec
JPEG Quality 80 Is Not a Setting
An AVIF Is an MP4 With One Frame
Every XML Feature Is One You Turn Off
Norway Is Not a Boolean
Your JSON Parser Disagrees With Mine
Nobody Agrees What a CSV Is
Word Documents Used to Be Filesystems
ZIP Files Are Read Backwards
Markdown Is Not a Format, It's an Argument
There Is No Such Thing as a Text File
How PNG Actually Stores Your Pixels
Deploying Hermes Agent With Ansible Without Creating a Snowflake
pnpm 11 Made the Safe Thing the Default
Boring Is a Feature
The Best Automation Has a Manual Escape Hatch
Cost and Latency Belong in the Score
Your Vibes Are Not an Agent Eval
Start With Ten Tasks You Actually Do
The Decision Log: A Lightweight Artifact for Agentic Coding
Run Your Whole Agent Stack on a $5 Box
If I Worked at Apple Again, I'd Build iApps
Your Agent Needs a Dollar Limit, Not a Token Budget
How I Would Build Observability for an Autonomous Agent
Your Justfile Is Your Repo's API
The Human Job Is Choosing What Not to Build
When I Use a Subagent and When I Don't
Where Should This Agent Knowledge Live?
Not Every Agent Task Needs an Issue
Tests Are Evidence, Not a Definition of Done
Every Feature Opens a Maintenance Account
Code Is Cheap Now. Decisions Aren't
Pi and Hermes Are Trying to Solve Different Problems
Don't Build a Full Agentic Harness. Wrap One Instead.
Your Coding Agent Should Not Own Your Editor
Graphify Turns Your Repos Into a Map You Can Query
What Are Embeddings, Really? (And Why You Can't Just Use GPT-4 to Make Them)
Why Your AI Stack Needs a Gateway
Autopilot for Your Docs: A Look at LangChain's OpenWiki
Amnesia-Free Agents: GBrain and the LLM Wiki Idea
Julia: High-Performance Computing Without the Two-Language Tax
Adding Types to JSON with Dhall
Building Speech and Thought Bubbles in Svelte
Running Mem0's Memory Backend as Real Infrastructure with Ansible
Geni vs. Goose for lightweight database migrations
Embeddings Are Cheap Enough for Personal Wikis Now
Build Your Own Skills Repo
Building a Linter for Your Obsidian Vault
The Capture Trap: Why Your Note Vault Is a Graveyard
Four Ways to Build Your Own Agentic Harness
Five Places RAG Shows Up in Agentic Systems
A Field Glossary for Agentic Knowledge Work
Two Kinds of Memory for Your CLI Agent
Karpathy's LLM Wiki: Your Second Brain, Maintained by the Machine
A Brief History of the Second Brain
Six Months with Git-Native Issue Tracking: Am I Still Using Beads?
How Do You Actually Review Code an Agent Wrote?
The Agentic Dev Space Is Moving Fast, and I'm Having a Blast
SvelteKit vs Astro: How I Choose in 2026
Why Supacode Became My Daily Driver for Claude Code
Day 19: Leap Seconds or When a Minute Has 61
Day 18: DST and The Related Software Bugs
Day 17 — Time Zones Are a Nightmare
Day 16: How the World Agreed on a Date Format (Except the US)
Day 15: The Man Who Synchronized the World
Day 14: The Bug That Didn't End the World, and the One That Still Might
Day 13: Unix Time, 1,780,620,532
Day 10: The Zero Point
Running Terraform in Your Existing CI Pipeline
HCP Terraform's Per-Resource Pricing Is a Trap
When You Should Skip Terraform Entirely
OpenTofu Is the No-Regrets Default for 2026 Infrastructure
A Dotfiles Manager That Snapshots Every Change
Your AI Coding Agent Can Read Every Secret on Your Machine
Buying Supply Chain Security in 2026: A Vendor Map
Sandboxing AI Agents Without Buying Anything
Python and Rust Have the Same Supply Chain Problem as NPM
Your Software Is Mostly Strangers' Code
Five Modern JavaScript Features That Make the Old Patterns Look Silly
Three CSS Features That Finally Let Us Delete the JavaScript
CSS Finally Got Inline Conditionals
JavaScript Finally Gets a Real Date API
Semantic Docs Spring Update: Astro 6, Auto-Releases, npm
SAST vs AI PR Review: Two Tools, Different Jobs
npmx.dev Is the NPM Frontend We've Been Asking For
Your Data Lake's Vulnerability Problem Is Really an Identity Problem
Your Data Lake Has a Permissions Problem
The Real Cost of Your Data Lake (It's Not the Storage)
AI Code Reviewers Won't Save You
Leading Teams When AI Does the Typing
What Is a Runbook and Why Should You Care?
How to Pick an Embedding Model (Without Overthinking It)
Pandoc vs MarkItDown: Two Tools, Two Eras
What Temporal Actually Does (And Why You'd Want It)
pgvector vs sqlite-vec: You Probably Don't Need Postgres
pgvector vs Pinecone: You Probably Don't Need a Separate Vector Database
How Kong Actually Works in Kubernetes
LangChain and LLM Routers, the Short Version
Your Brain vs. a Large Language Model
What I Learned Building My First Chrome Extension
NotebookLM Is Just RAG With a Nice UI
AI-Assisted vs AI-Agentic Coding
Claude Opus 4.7 Is Here
Agentic Development Trends: What's Changed in Early 2026
Why javascript:void(0) Needs to Stay in the Past
The Human's Guide to the Command Line: From Script to System
How Active Noise Canceling Actually Works
What Is an AI Agent, Actually?
SvelteKit Forms: Back to Basics (And Why That's a Good Thing)
Routing in SvelteKit vs Next.js vs Astro
Svelte 5 Runes: A React Developer's Guide to Reactivity
Svelte vs React: State Management Without the Ceremony
Svelte vs React: The Virtual DOM Tax You Might Not Need
Lisette is a New Rust-to-Go Language, So I Built It a Test Library
Hiding Poems Inside Images
The Human's Guide to the Command Line: Your First CLI App
The Human's Guide to the Command Line (macOS Edition)
What Companies Are Actually Paying for Application Security
I Benchmarked JSON Parsing in Bun, Node, Rust, and Go
I Wrote Multiple CUE Parsers and Benchmarked Them Against JSON
Is There Something Better Than JSON?
Multi-Repos Are Underrated
CLI First, GUI Never
The Death of Clever Code
Your AI Agent Needs a Task Manager
Using Claude to Think Through a Space Elevator
Just: The Command Runner
From SUnit to Vitest: A Brief History of JavaScript Testing
Garbage Collection: How Python, JavaScript, and Go Clean Up After Themselves
I'm Just Going to Use Zed
Voice-to-Text in 2026: The Tools and Models Worth Knowing About
Everything Is Eventually a Database Problem
Your Context Window Is a Budget — Here's How to Stop Blowing It
When to Use Python Over Bash
Local Secrets Manager - Dotenv Encrypter
Why Ghostty Is My Terminal for Agentic Work
How to Write a Good CLAUDE.md File
Claude Code Skills vs Plugins: What's the Difference?
Claude Code Now Has Two Different Security Review Tools
Managing Your Context Window in Claude Code
Claude Code Prompts for Taming Your GitHub Repository Sprawl
What I Actually Mean When I Say "Vibe Coding"
Stop Using pip. Seriously.
Why Testing Matters
Serverless and Edge Computing: A Practical Guide
Defending Your Node Modules: Security Tools and When to Rewrite Dependencies
Ditch Jest for Vitest: A Ready-Made Migration Prompt
JavaScript Package Managers: NPM, Yarn, PNPM, and Bun Compared
Why Svelte 5 Wants `let` Instead of `const` (And Why Your Linter Is Confused)
REPL-Driven Development Is Back (Thanks to AI)
Switching to mise for Local Dev Tool Management
Knowledge Without a Knower
The Broken Promise of Reach
First Impressions of OpenAI's Codex App
Why IndiePub Matters More Than Ever
Why Data Modeling Matters When Building with AI
JavaScript Still Doesn't Have Types (And That's Probably Fine)
2026: The Year We Stop Blaming the Tools
The Rise of Spec-Driven Development: A Guide to Building with AI
AMP Code: First Impressions of a Claude Code Competitor
Teaching AI to Hide Messages in Plain Sight
Two Changes in Claude Code That Actually Matter
Security and Reliability in AI-Assisted Development
Learning to Program in 2026
Two Arguments Against AI in Programming (And Why I'm Not Convinced)
Claude Cowork: First Impressions (From the Sidelines)
Beads: Git-Native Issue Tracking for AI-Assisted Development
36 Framework Fixtures in One Session: How Beads + Claude Code Changed Our Testing Game
Why Airtable Killed Google Sheets for Me
The Markdown Mode Manifesto
Twenty Years of DevOps: What's Changed and What Hasn't
Introducing api2spec: Generate OpenAPI Specs from Source Code
198 posts in the archive
START
llbbl.exeprojects/posts/experiments/subscribe.dlg
© 2026v1.0.0