Written from real work,
not a pitch

Reliability

Lighthouse score is not a design review

~7 min read · 31 Jul 2026The sweep passed on every axis and a real reader still said it looked like a toy. The problem was the set we chose to measure, not that the thing has no metric.AI · Framing the ask

The freshest work in your head will crowd out the next answer

~6 min · Jul 31, 2026Context engineering is deciding what belongs in the context a model sees, not just the sentence you type. Where it differs from prompt engineering. Reliability · Testing

The Passing Test Was Telling Us to Delete Real Customers

~5 min · Jul 30, 2026One check said the signups table must hold zero rows. Six people had signed up. The quickest way to make it pass again was to delete all six.Accounting · Tax

There Are Two Profit Numbers, and Tax Uses the One Not in Your Books

~7 min · Jul 30, 2026Same one million baht. Accounting says expense it, tax says add it all back. Both are right, because the two numbers exist for different jobs.Accounting · Review

Preparing Statements ≠ Reviewing Them: The Skill AI Cannot Replace

~8 min · Jul 30, 2026A preparer asks did I record everything. A reviewer asks does this number make sense. The second question is the part you cannot hand to AI.AI · Tool comparison

Two Ways to Let AI Drive a Browser, Tested for Real

~8 min · Jul 29, 2026ego-browser vs BrowserClaw, tested on the same task: two architectures of the AI agent browser, and which to pick.AI · Repo Review

Pool three idle machines into one LLM endpoint you own

~7 min read · Jul 25, 2026A review of mesh-llm, an open-source tool that pools GPUs across machines into one OpenAI-compatible LLM endpoint, from deploying it for real across three machines.AI · Repo Review

Andrew Ng shipped an AI coworker that works on your machine, and still lets you keep control

~8 min read · Jul 24, 2026A review of openworker, Andrew Ng's local-first AI coworker, from reading the whole codebase. The idea worth borrowing is how it keeps AI controllable at the tool-call boundary.AI Infrastructure · Cost

The GPU budget doesn't leak under load. It leaks when jobs crash silently

~7 min read · Jul 24, 2026Serverless GPU bills by the second and that part delivers. But a near-6-minute cold start forces an async design, and the budget leaks when jobs break and the CLI never shows it.AI · Reliability

The daemon that died for two weeks, and nobody noticed

~9 min read · Jul 24, 2026A daemon runs in the background all the time. Its real danger isn't crashing, it's crashing with no signal. Told through one that died silently for two weeks, and how to make yours make a sound.AI · Design System

AI Built a Convincing Design System, and Got It All Wrong

~4 min · Jul 16, 2026Asked AI for a design system and it picked a tidy, confident palette that was wrong three times over. The rule that fixed it: reflect what exists, do not invent.Cost & Models

The cheap model that gets to be wrong beats the expensive one that can't

~6 min read · Jul 15, 2026Six hard bugs, measured: a cheaper model allowed to loop 3 times matched the expensive one-shot at 1x cost against 3.76x, because a checker to fire against lets it be right cheaply.AI Production · Fable 5

Why Claude Fable 5 switches models mid-conversation, and what it costs

~4 min · 11 Jul 2026Fable 5 hands risky requests to Opus 4.8 automatically. When a normal task gets switched anyway, that is a false alarm on the system side.AI Agents

AI writes the code, but who signs for it

~3 min · 10 Jul 2026As AI writes more of the code, the missing thing is accountability. The day it became a real file across 5 repos.SME · Inventory

Your cash didn't disappear. It's asleep in the stockroom.

~10 min read · Jul 10, 2026Inventory analysis with a handful of numbers: days of inventory, ABC-XYZ, and a cut score, straight from the spreadsheet you already have.Accounting · Data Accuracy

The model misread a bill by five hundred baht. What caught it?

~6 min read · Jul 10, 2026One bill, two confident answers. How an arithmetic verification layer catches misreads on its own.AI Engineering · Reliability

Stop begging the model for JSON. Constrain it with a schema

~6 min read · Jul 10, 2026Three failed runs of a model that promised JSON and rambled instead, until we moved enforcement to the decoding layer.AI Infrastructure · Cost

Modal GPU on a budget: 7 submerged stumps before the first clean run

~8 min read · Jul 10, 2026Pay-per-second GPU, every stump we hit, and real per-bill costs from our logs.Accounting · Tax

Certificate in lieu of receipt: when and how to use it for Thai tax deductions

~8 min read · Jul 10, 2026Taxi fares and market purchases come with no receipt. The Revenue Department already has a path for them.Accounting

Choose an AI Bill Tool by the Job, Not the Feature List

~7 min · Jul 9, 2026Parnuan, Paypers, FlowAccount, ask ChatGPT/Claude, or DIY each takes your bill somewhere different. How to choose from real accounting work.Product & PM

What is spec-driven development, and why non-coders can do it

~6 min read · Jul 9, 2026It's all over AI coding and sounds like a programmer thing, but the heart of it is thinking clearly about what to build, which non-coders can do.Product & PM

How to write a PRD your AI can build from (template + example)

~8 min read · Jul 9, 2026You've been told to write a spec first, but the blank page is the hard part. An 8-section template and a real worked example.Product & PM

A vibe is not a spec: write a PRD your AI can build from

~7 min read · Jul 9, 2026Vibe coding is fast, but without a spec it builds the wrong thing. Here's the minimal PRD that makes AI build right.Reliability

Build your own social listening tool in a weekend

~9 min read · Jul 8, 2026Wire scrapers to an AI and listen to social yourself in one weekend, plus the five things you must get right.Dev Setup

Setting up Ghostty + Zsh without the one-command installer

~6 min · 8 Jul 2026A curl-into-bash script can rewrite your whole shell. How to build it piece by piece, safely.Dev Setup

Ghostty vs iTerm2 vs Terminal.app: the best Mac terminal

~6 min · 8 Jul 2026All three compared from real use: speed, config, and which fits working with AI agents.Reliability

The safety filter that was silently deleting your data

~9 min · 8 Jul 2026Gemini's safety filter flagged ordinary content as sensitive and dropped records with no error. The fix: make the extraction step a swappable engine.Knowledge & Memory

AI knowledge management: build a system you can trust, layer by layer

~8 min · 8 Jul 2026How to assemble an AI knowledge management system layer by layer, each one earning trust before it is believed.Production Guardrails

allowedTools isn't file access: why your agent still gets permission denied

~6 min read · Jul 8, 2026allowedTools allowed Read, yet the agent can't read outside cwd. Glob finds it, Read is denied, because permission and --add-dir are separate layers.SEO & Discovery

Google search favicon: why it shows a globe instead of your icon

~6 min read · Jul 8, 2026Google shows a globe instead of your favicon when Googlebot can't fetch it. Diagnose with curl, fix at the root.Funnel & Content

Gated content without a backend: email-unlock on a Cloudflare Worker

~8 min read · Jul 7, 2026Gate content on a fully static site with no backend or HubSpot: crawlable teaser, worker-served payload after unlock, plus the origin trap that nearly left the gate open.Skills & Plugins

Your AI agents should not all carry the same skills

~8 min read · Jul 6, 2026Six agents, one afternoon, three misses: names that lied, a layer that failed silently, and versions shifting under our feet. Ends with a lint that compares the list against disk. Claude Skills

“Make it pretty” is the wrong prompt

~8 min read · Jul 6, 2026Every AI-designed site looks the same. A design skill that refuses to code until three strategy questions are answered, told through a real admin-page before/afterReliability

We Told the Agent to “Verify It Works”. It Picked the Easiest Possible Meaning

~5 min read · Jul 6, 2026A real migration night: every line of the report was green and the system was down. How one runnable acceptance line closes the gap, plus the three enforcement layers that end with proving the checker itself.AI Agents

We Interviewed the Most Expensive AI Model in the House About Making Cheaper Ones Do Its Job

~12 min read · Jul 6, 2026A real interview with Anthropic’s new top tier: the principles it was trained on, the weaknesses it volunteered, what it is not better at, and six levers that let Opus stand in when costs must come down.Cost & Models

Claude Opus vs Sonnet vs Fable 5: Which Model for Which Work

~7 min read · Jul 5, 2026The night our priciest model burned its credit mid-task rewrote our fleet rulebook: the criterion is ambiguity, not difficulty, plus six levers that squeeze a mid tier above its weight.Cost & Model

23 free models on OpenRouter: the quota is real, the queue is not yours

~8 min read · Jul 4, 2026Tested with a real key: the free llama returned 429 nine times in a row with a full quota. A verdict table of which jobs fit which kind of free.Cost & Model

Your static site can have AI. No backend. No API key.

~9 min read · Jul 4, 2026Add an AI box with Cloudflare Workers AI, with real measured numbers: 124 neurons per call, ~80 free calls a day, and a two-layer cap that keeps the bill at zero by proof.Cost & Model

Your AI agent doesn't need a bigger machine, it needs a home that never turns off

~10 min read · Jul 4, 2026We moved a full agent fleet to a $14/month VPS and measured for three weeks: 3GB of RAM, not 8GB, and CPU runs out first. With the real invoice and a live three-provider price table.Cost & Model

A 27B model on a rented GPU with vLLM: the traps nobody writes down

~9 min read · Jul 3, 2026The hands-on half of the GPU-rental story: the busy empty machine, config in two places, what survives a stop, and the tunnel that must reconnect itself.Cost & Model

Renting a GPU for your own LLM: the expensive part is the idle hours

~9 min read · Jul 3, 2026Per-token APIs vs an hourly GPU vs serverless GPU: live prices from three providers, a real break-even formula, and the data line the price tag never shows.AI Agents

Claude Fable 5 as the head, everything else as hands

~15 min read · Jul 3, 2026The most expensive model as team lead, four cheaper subagents as hands, one real migration assessment. What worked and what broke.Claude Skills

Claude Code hooks: a done-voice

~8 min read · Jul 3, 2026Send an agent on a long run and walk away. A hook speaks what finished, and which tab.Product & PM

AI product management: what's left for PMs when the frameworks become skills

~8 min read · Jul 2, 2026The pillar tying four repo reviews together. When the field's frameworks are a command away, knowledge is cheap and judgment is the scarce part.Product & PM

The whole GTM playbook in 12 skills, wired to cascade

~7 min read · Jul 2, 2026A review of Maja Voje's GTM Strategist skills. The win is the design, set product context once and each phase feeds the next, not a longer framework list.Product & PM

The YC president open-sourced his own stack. It's about taste

~7 min read · Jul 2, 2026A review of Garry Tan's gstack. Multiple review lenses encode taste into software, and AI models recommend while users decide.Product & PM

Product discovery as a pipeline, with two judgment calls baked in

~6 min read · Jul 2, 2026A review of Else van der Berg's product-discovery-skills. Teresa Torres's method, with misfits as signal and importance over prevalence.Product & PM

The whole PM craft as ~68 skills

~6 min read · Jul 2, 2026A review of Pawel Huryn's pm-skills. Roughly 68 skills across 9 plugins, and the intended-vs-implemented doc-vs-code audit that stands out.Claude Skills

9arm's skills repo is small, but it has one cost idea worth stealing

~5 min · Jul 2, 2026Don't burn your expensive model on grunt work. Hand it to a cheap one, keep the expensive model for judgment. The sharp cost idea inside 9arm's small repo.Claude Skills

Addy Osmani agent-skills review: 24 production skills, 72.6k stars

~6 min · Jul 2, 2026Addy Osmani, a Google web engineer, opened a 24-skill repo covering the full lifecycle. The standout the other three lack: browser testing and performance. The closer of the review series.Claude Skills

mattpocock/skills review: a real engineer's .claude, 160k stars

~7 min · Jul 2, 2026Matt Pocock opened up his own .claude folder: ~36 skills you fork and remix. The third in the series reviewing trending skill repos, after superpowers and karpathy.Claude Skills

Karpathy Skills review: one CLAUDE.md, 189k stars

~7 min · Jul 2, 2026A 189k-star repo that distills how AI breaks when it codes into 4 principles in a single CLAUDE.md. More minimal than superpowers, and how to pick between them.Claude Skills

Superpowers review: the AI agent repo blowing up, 243k stars

~9 min · Jul 2, 2026The AI coding-agent repo everyone's installing. I tried it and read the source: is it worth it, what to take, and the core idea that matched a skill I had already built.AI Agents

What is an AI agent, vs a chatbot?

~6 min · Jul 2, 2026If you open ChatGPT, ask, and copy the answer out, that is a chatbot, not an agent. Here is the line, and the on-ramp to the 7-layer series.AI Agents

AI Agent Architecture: The 7 Layers a Production Agent Needs

~9 min read · 1 Jul 2026Search "ai agent architecture" and you get textbook diagrams. Here are the 7 real layers that separate a production agent from a demo.Reliability

The Ported Code Passed Every Test. It Was Still Wrong.

~9 min read · 1 Jul 2026Self-written tests only prove the code matches what you think. Characterization testing uses the original that already runs as the answer key, diffed line by line.Accounting

From a Ledger Entry to Its Source Document, Checkable

~7 min · Jul 1, 2026Vouching proves every entry has a real document behind it. Let AI match entries to documents, then surface the ones with none for a person to decide.Accounting

From Statement and Books to a Reconciliation You Can Check

~7 min · Jul 1, 2026A good reconciliation's output isn't the word "matched". It's every unmatched row with a reason. Let AI match, then classify what didn't.Accounting

From a Photo of an Invoice to Import-Ready Data

~7 min · Jul 1, 2026The problem isn't that AI can't read your invoices. It's that you get text you still can't import. Here's how to get import-ready data, with the numbers matching the source.Accounting

Before AI helps you file, read the map first

~5 min · Jul 1, 2026A full walkthrough of DBD e-Filing, then a task analysis of where AI helps and where a person must decide.Claude Skills

A Skill Isn't Just a Longer Prompt

~5 min · Jun 30, 2026What a Claude Code skill is, how it differs from a prompt, and when to stop re-typing and package one the model reaches for on its own. Claude Skills

Don't Let the LLM Do the Math

~8 min · Jun 30, 2026A reliable skill isn't the LLM doing everything. Work that must be exact every run is test-locked code; only what needs judgment goes to the model. Reliability

The tool you trust to catch mistakes was the one making them

~7 min · Jun 30, 2026Sometimes the thing you use to catch mistakes fails silently, then reports "no bugs" while it checked nothing. The way to catch it: make it fail once, on purpose, with mutation testing and a positive control. AI Agents

I ran three AI coders at once. The agents weren't the cost.

~5 min · Jun 30, 2026Three coders in parallel and the bill barely moved. The expensive part is human time and the defects that slip through. The shape we use: coders on a subscription, a separate reviewer, and a human holding merge. Knowledge & Memory

Your agent has all its rules. That doesn't mean it reaches them in time

~6 min · Jun 26, 2026When your rules file grows too big to load, what do you cut? The deeper question is why a rule you already stored sometimes doesn't fire when it should. Reliability

When several AIs agree, that isn't proof

~4 min · Jun 26, 2026Pulling in several AIs to review your work feels more thorough. But the moment they all agree might be the most dangerous one. AI Agents

Hand work to your AI fleet through a board, not a chat

~5 min · Jun 26, 2026We had AI agents that finish work on their own, but chatting with each one made us the bottleneck. Put the work on your Linear board, tag it for the fleet, and the result comes back on the card. Irreversible work waits for a human. AI Agents

When parallel AI agents corrupt each other's work

~5 min · Jun 26, 2026Let parallel AI agents write in one shared directory and they trample each other. The fix is structural: one git worktree per task, created by the orchestrator. AI Agents

Don't marry one model: separate the soul, engine, and orchestrator

~6 min · Jun 26, 2026The day you want a cheaper model, you shouldn't have to rebuild the agent. Separate three layers up front and swapping models becomes a config change. Knowledge & Memory

AGENTS.md and SOUL.md: sharing knowledge across an AI fleet

~5 min · Jun 26, 2026We run several AI models as a team, but each one's knowledge stays locked in its own format. The fix isn't a big database. It's two open files every model can read. AI Agents

Connect two AI agents across machines without opening a port

~7 min · Jun 26, 2026The home box is behind NAT and the cloud box has every port closed for safety, yet two AI agents still talk across machines, through an SSH tunnel that already ships on every computer. Privacy & Security

The data that should never flow up

~5 min · Jun 25, 2026A three-layer defense for data and AI. Cloud knowledge flows down to help; your secrets never flow back up. Plus the three sneakiest leak points. Cost & Models

The best model is often the wrong one

~5 min · Jun 24, 2026Choose a local LLM by the job, not the benchmark. Qwen vs Gemma from real work, plus QAT, MTP, dense vs MoE. Reliability

Let Codex review the code AI wrote

~7 min · Jun 23, 2026The writer and the reviewer shouldn't be the same model. Two times a second model caught a bug we'd missed in our own code. Knowledge & Memory

The real memory stack we run: Graphiti + FalkorDB

~6 min · Jun 23, 2026Behind "give the AI memory" are five wired layers and two bugs that fail silently, plus how to check your agent really remembers. Knowledge & Memory

Design AI memory the way human memory works

~9 min · Jun 23, 2026Memory that remembers is not about storing, it is about pulling the right thing back. Design it in layers like human memory, with the real tools we run. AI Agents

AI coding agent fleet: the Kanban swarm pattern

~8 min · Jun 21, 2026One AI coder is easy. The pattern that lets a lead split a goal into cards, run workers in parallel in separate worktrees, and gate the merge with a second-engine review. AI Agents

I Let AI Post to Social for Me, and the 3 Guardrails That Keep It Safe

~7 min · Jun 21, 2026Auto-posting to Facebook and LinkedIn every day: the step-by-step setup, and the guardrails that keep an irreversible public action from going wrong. Reliability

Why Your AI Agent Lies to You

~6 min · Jun 20, 2026AI doesn't lie at random. It guesses what's plausible and says it with confidence, and it fools you best when it reports "done." AI Visibility

Make Your Site Agent-Ready: I Did 8, Refused 3

~7 min · Jun 19, 2026I ran my site through isitagentready. Do the real checks, refuse the ones that advertise capabilities you do not have. The missing checkmarks are deliberate. Privacy & Security

Let Claude cite you, block the trainers

~7 min read · Jun 18, 2026Blocking every AI bot feels safe, but it also shuts the door that lets people find you through AI. How to allow them one at a time on Cloudflare. Cost & Models

50 MCP servers, six calls

~5 min read · Jun 18, 2026A tool you connect but never call taxes the context window every turn. How to audit and prune. Claude Skills

Evaluating Claude Code plugins

~5 min read · Jun 17, 2026ponytail and headroom trended together. The test for whether to install. Cost & Models

You Can Still Run Agents on a Subscription (For Now)

~8 min read · Jun 16, 2026Anthropic announced a billing split then paused it. Right now agents still run on the plan, but the door may close. AI Visibility

Your Site Is Live. But Who Can See It?

~6 min read · Jun 16, 2026Live does not mean found. 5 steps a one-person business can do in an afternoon so both Google and AI see your site. AI Agents

Not Every Action Needs a Human

~8 min · 15 Jun 2026Human in the loop is treated like a light switch. Here's the third option: tier the loop, and decide who decides per action. Knowledge & Memory

Systematize turning speech into a second brain you can trust

~8 min · 14 Jun 2026Doing it by hand works for one clip, but dozens need a system. Here is the four-stage pipeline that makes the evidence gate run on its own. Knowledge & Memory

Transcribe on your own machine, faster than Whisper

~9 min read · 14 Jun 2026A batch of Thai lecture audio would not finish transcribing all day, until I learned the slow part was the model architecture, not the hardware. Here is the transducer fix. AI Agents

Bring your AI into Discord, without handing over the keys

~11 min · 14 Jun 2026Connect AI to Discord so you can chat from anywhere, with a gate that controls who can command the bot, and keeps a chat message from becoming a command. Knowledge & Memory

Turn speech into trustworthy notes, without letting AI make things up

~10 min · 14 Jun 2026Summarize meetings or lectures into knowledge notes while keeping AI from adding things no one said, then store them so they stay findable.