Costs & subscriptions · Tools & costs
Eight Hours a Day Coding Against AI: Relay Plans vs. Official Subscriptions, Which Adds Up Better
Running the numbers for 8 hours a day: relay plan quotas, official subscription pricing, and the hidden costs folded into the price.
Latest
RSS ↗
The Real Cost of Long Context: How to Do the Token Math
Context isn't free. The same conversation, placed at the start instead of the end, can cost more than twice as much.
03
The Default Language on a Multilingual Site: Why It Shouldn't Carry a Language Prefix
Default language at the root, others prefixed — the simplest hreflang setup. URL design, x-default, and the classic "can't switch language" bug.
04
What one call costs: three variables in the token ledger
Input, output, and cache hits — three variables decide the real cost of a single request. The formula worked through with concrete numbers.
05
Four ways to schedule tasks: what's beyond cron
cron is simple but hard to monitor. systemd timers, container schedulers and in-app scheduling each trade off differently.
06
GA4's data model: why events matter more than pages
GA4 replaced pageviews with events plus parameters. Without learning that language, your reports will never reconcile.
07
Turning Repetitive Operations into Skills: Which Workflows Are Worth Keeping
Not every operation deserves to become a skill. There is only one criterion: will it repeat more than three times within a month?
08
How to Write CLAUDE.md: Handing Project Common Sense to the Model
It's not documentation — it's context reread every session. Writing too much is the same as writing nothing.
09
Writing an MCP Server: Four Constraints for Tool Design
More tools make the model dumber. Four constraints to keep your server "few but precise" instead of "complete but unused".
10
Does knowledge base search need vectors? Start with full-text, don't rush to embeddings
From a dozen notes to a few thousand, full-text search is enough. A real knowledge base MCP setup, and when to switch to vector search.
11
An AI Image and Video Asset Pipeline: From Prompts to Batch Output
Generating dozens of images makes one-by-one chat prompting unworkable. Split prompts, naming, filtering and archiving into a repeatable pipeline.
12
vLLM, llama.cpp, Ollama: Where Each Inference Stack Fits
Throughput, ease of use and quantization support: you can't have all three. Pick by use case, not by benchmark scores.
13
Running LLMs on a 4090: The Real Limits of 24 GB of VRAM
Whether a 70B fits, how far to quantize, how much concurrency breaks it — judged by the VRAM budget, not the parameter count.
14
The Three Layers of AI Coding Tools: Completion, Chat, Agents — Don't Buy the Same Layer Twice
Completion, chat, and agent tools solve different problems. Knowing which layer you're missing saves more money than subscribing to five tools.
15
Stripe vs PayPal: The Real Barriers to Collecting Payments as an Indie Going Global
Rates are only surface-level. Account approval difficulty, chargeback handling, and payout timelines decide whether you can use it long term.
16
Tokyo, Singapore, Frankfurt: Latency and Price for Cloud Servers
The same specs can cost twice as much across three regions. What latency, bandwidth and compliance each buy you.
17
Three Monetization Paths for Content Sites and Their Startup Thresholds
Ads, subscriptions, referral and tool payments: three paths with completely different startup thresholds, revenue curves and failure modes.
18
Is a new site worth building? First calculate the "verifiable minimum investment
Before writing the first line of code, use five questions to judge whether the direction is right. This matrix cut most of the sites I originally planned to build.
19
Structured Data: Which Schemas Actually Earn Clicks
Article, FAQ, Breadcrumb, Product — not all are worth adding. Which still show in search results, and which are already gone.
20
The Three Core Web Vitals Metrics, and How to Actually Bring Them Down
LCP, INP and CLS each have clear causes. Find which one is failing first, then optimize — don't just throw "performance plugins" at it.
21
Site Structure: A Minimal Design for Categories, Tags, and Internal Links
Categories are for readers, tags are for search, internal links are for crawlers. Design all three at once and it will end up a mess.
22
Beyond Keyword Tools: Reverse-Engineering Topic Ideas from Search Console
Keyword tools tell you what others search for; Search Console tells you what you already have. The latter is worth more, and fewer people use it.
23
Five AdSense Rejection Reasons and How to Fix Each
Low value content and unclear navigation are completely different problems. Five rejection categories, each with concrete fixes.
24
AdSense Keyword Site SOP: Eight Steps — Get Approved Before You Add Features
A full workflow from keyword research to matrix review. The core rule: before you have traffic, keep features minimal and get approved first.
Topics
IndexAI tools
Models, CLIs and editors: which layer is worth paying for.
4 notes 02Costs & subscriptions
Subscriptions, relays and metered billing: the actual bills.
6 notes 03Building & SEO
From keyword to approval, then to traffic.
8 notes 04Workflows & MCP
Turning repeated actions into something callable.
6 notesNo newsletter needed — use a reader.
No paywall here, and no marketing mail. For steady updates, add the RSS feed to your reader; it beats waiting on push.