How I Turned One Successful Run Into a Reusable Codex Skill
How I Turned One Successful Run Into a Reusable Codex Skill

How I turned a successful Codex run into a reusable skill using a short SKILL.md, a detailed workflow reference, and a deterministic PowerShell script.

September 1, 2026 Read
How I Built a NotebookLM Podcast With an Audience of One
How I Built a NotebookLM Podcast With an Audience of One

I use Codex and NotebookLM to research a topic, generate two Audio Overviews, publish them through a self-hosted RSS feed, and receive them in BeyondPod.

September 1, 2026 Read
I Was Wrong About Project Glasswing
I Was Wrong About Project Glasswing

OpenAI's Hugging Face breach, Anthropic's Mythos 5 incidents, and UK AISI's findings changed my view of Project Glasswing and AI agent containment.

August 20, 2026 Read
Codex vs Claude Code: My Honest Comparison
Codex vs Claude Code: My Honest Comparison

Codex vs Claude Code, from a developer who lives in these tools. Codex leans on cloud task delegation; Claude Code on a local terminal agent. How to choose.

May 26, 2026 Read
Claude Code vs Cursor: My Honest Comparison
Claude Code vs Cursor: My Honest Comparison

Claude Code vs Cursor, from a developer who uses both. Claude Code is an autonomous terminal agent; Cursor is an AI-first editor. Here is how to choose.

May 25, 2026 Read
The Long Tail of Bespoke Software: What Gets Built When the Marginal Cost of Code Drops
The Long Tail of Bespoke Software: What Gets Built When the Marginal Cost of Code Drops

When AI agents drop the marginal cost of writing code to near zero, the most interesting first-order effect is bespoke automation that was never economically viable before. Here is what that looks like in practice.

May 5, 2026 Read
Who Wrote This Code? Authorship and the Specification Crisis in AI-Assisted Development
Who Wrote This Code? Authorship and the Specification Crisis in AI-Assisted Development

When a 200-word prompt produces 500 lines of code, who 'wrote' it? Code review becomes the new authoring, and software engineering's specification crisis is finally visible.

May 4, 2026 Read
A Taxonomy of Claude Code Prompt Shapes (and the Anti-Patterns That Waste Tokens)
A Taxonomy of Claude Code Prompt Shapes (and the Anti-Patterns That Waste Tokens)

Six Claude Code prompt shapes that work, plus the anti-patterns that waste tokens. Generate, refactor, debug, explain, orchestrate, review, with one real example of each.

May 3, 2026 Read
Is Claude Code a 5th-Generation Language? What "Just Describe What You Want" Really Means in 2026
Is Claude Code a 5th-Generation Language? What "Just Describe What You Want" Really Means in 2026

Claude Code matches the textbook 5GL definition: describe what you want, the system figures out how. The agentic layer on top is what is new in 2026.

May 1, 2026 Read
How I Built a Self-Hosted Personal Finance Dashboard With Claude Code
How I Built a Self-Hosted Personal Finance Dashboard With Claude Code

I built a self-hosted personal finance dashboard in 12 days with Claude Code, React, Node.js, PostgreSQL, Tiller, and Docker. Here is the architecture, data pipeline, and what actually worked.

April 11, 2026 Read
Graphify Review: I Tried It on My Codebase With Claude Code
Graphify Review: I Tried It on My Codebase With Claude Code

Graphify builds a codebase knowledge graph for Claude Code. I tried it on my own project: the idea is real, the daily workflow is still rough.

April 11, 2026 Read
Anthropic Mythos: Why Everyone Is Freaking Out About Project Glasswing
Anthropic Mythos: Why Everyone Is Freaking Out About Project Glasswing

Anthropic's Claude Mythos Preview can reportedly find and exploit real vulnerabilities. Project Glasswing also shows how AI labs sell capability with danger.

April 11, 2026 Read