Claude Code, Codex, and Copilot: AI Coding Agents in 2026
· 9 min read · ai, agents
How the major coding agents differ in autonomy, environment access, and pricing — and how to pick one for your workflow.
Blog
Practical writing on AI, web analytics, link tracking, SEO, and building with Next.js and Supabase.
Topics: agents · ai · analytics · auth · http · llm · marketing · markets · nextjs · postgres · security · seo · supabase · tutorial
What Is RAG? Retrieval-Augmented Generation, Explained
· 10 min read · ai, llm
How retrieval-augmented generation works end to end — embeddings, vector search, chunking, and the failure modes the demos never mention.
Prompt Injection, Explained: Why LLM Apps Are Hard to Secure
· 9 min read · ai, security
Why models can’t reliably tell instructions from data, what the lethal trifecta means for agents, and the mitigations that actually help.
Fine-Tuning vs RAG vs Prompt Engineering: When to Use Which
· 9 min read · ai, llm
A decision guide for getting domain knowledge into an LLM app — costs, failure modes, and the order you should actually try things in.
What Is a Token? How LLMs Actually Read Text
· 8 min read · ai, llm
Why models see subwords instead of words, why code and non-English text cost more, and how token counts drive pricing, context, and speed.
How to Run an LLM Locally in 2026
· 10 min read · ai, llm
Ollama, llama.cpp, and what your hardware can actually handle — model sizes, quantization, and when local beats the API.
Circular AI Deals, Explained: When Vendors Finance Their Customers
· 9 min read · ai, markets
Chipmakers investing in the labs that buy their chips: how circular financing works, why it spooks investors, and the dot-com parallel.
Next.js SEO: A Practical Checklist for App Router Sites
· 10 min read · nextjs, seo
Metadata, sitemaps, structured data, OG images, and rendering choices — everything an App Router site needs to be properly crawlable.
Claude, GPT, or Gemini: How to Choose an LLM in 2026
· 8 min read · ai, llm
Capability tiers, context windows, and pricing across the big three model families — and the questions that actually decide which one fits your product.
LLM API Pricing, Explained: Tokens, Context, and Caching
· 10 min read · ai, llm
How per-token pricing actually works, why output tokens cost more, and how prompt caching and batch processing cut real bills.
What Are AI Agents, Actually? A No-Hype Explainer
· 9 min read · ai, agents
An agent is a model in a loop with tools. What that means concretely, when agents beat single prompts, and where they still fail.
What Is MCP? The Model Context Protocol, Explained
· 10 min read · ai, agents
Why AI tools converged on one protocol for connecting models to data and tools, how MCP servers and clients work, and when to build your own.
AI Capex and the Stock Market: What the Buildout Numbers Say
· 9 min read · ai, markets
Hyperscaler capital spending, AI’s share of index gains, and how to think about concentration risk — an explainer, not investment advice.
UTM Parameters, Explained: A Practical Guide to Campaign Tracking
· 10 min read · analytics, marketing
What utm_source, utm_medium, and utm_campaign actually do, how they flow into GA4, and the naming conventions that keep campaign reports clean.
301 vs 302 Redirects (and 307, 308): Which One Should You Use?
· 8 min read · http, seo
How each HTTP redirect status code behaves, what it means for SEO and caching, and simple rules for choosing the right one.
Build a URL Shortener with Next.js and Supabase
· 10 min read · nextjs, supabase, tutorial
A complete walkthrough of building a personal link shortener: schema design, redirect handling, click analytics, and deployment on Vercel.
What Is GCLID? How Google Ads Click Tracking Really Works
· 9 min read · analytics, marketing
Where the gclid parameter comes from, how offline conversion imports use it, and why stripping it breaks Google Ads attribution.
Do URL Shorteners Hurt SEO? What Actually Happens to Link Equity
· 9 min read · seo, http
How search engines treat shortened links, when a 301 passes PageRank, and the cases where a shortener genuinely costs you.
Supabase Row Level Security: Practical Patterns That Scale
· 10 min read · supabase, postgres
RLS policies for common app shapes — per-user data, admin service roles, public read tables — with the pitfalls that cause silent empty results.
From middleware.ts to proxy.ts: Migrating Middleware in Next.js 16
· 8 min read · nextjs
Next.js 16 renames middleware to proxy. What changed, what stayed the same, and a step-by-step migration with auth gating as the example.
Magic Link Auth in Next.js with Supabase: The Complete Setup
· 10 min read · supabase, nextjs, auth
Passwordless email sign-in with @supabase/ssr — auth callback routes, session refresh, protected routes, and the redirect gotchas.