Blog

Tutorials and field notes on encoding what you already know into agent workflows: hooks, commands, evals, and the guardrails that make them safe to run.

40 posts

Run Claude Code From Your Phone with tmux and Tailscale
Tutorials & How-Tos7 min read

Run Claude Code From Your Phone with tmux and Tailscale

A claw that props your laptop lid open solves the wrong problem. Put your agents on a headless Mac mini and reach them from your phone with tmux and Tailscale.

Measure AI Cost Per Feature, Not Per Pull Request
Problem & Solution8 min read

Measure AI Cost Per Feature, Not Per Pull Request

Usage-based AI pricing breaks fixed-cost tool budgeting. Measure agent spend against the business feature you shipped, not the seat, the token, or the PR.

The Model Race Is Cooling. Enterprise Is the New Race.
Opinions & Analysis8 min read

The Model Race Is Cooling. Enterprise Is the New Race.

AI labs are shifting budget from frontier model tiers to enterprise products and expertise. That pivot is the best thing to happen to model portability.

The Race to Own the Production Agent Layer
Comparisons & Reviews7 min read

The Race to Own the Production Agent Layer

Twelve managed-agent platforms, one architecture. Why the race to run agents in production already has a blueprint, and how to bet on the shape, not the vendor.

Managed Agents: The Production Layer Above Coding CLIs
Guides & Fundamentals6 min read

Managed Agents: The Production Layer Above Coding CLIs

Managed agents are the 2026 story. What they are, why they differ from CLIs like Claude Code and Codex, and a code-first POC built on Anthropic's platform.

Build Your AI Harness on Marketplace Plugins, Not Vendors
Guides & Fundamentals8 min read

Build Your AI Harness on Marketplace Plugins, Not Vendors

Marketplace plugins ship five portable primitives — commands, skills, sub-agents, hooks, MCPs — that let you build business AI workflows across vendors.

Where to Invest in AI Without Getting Vendor-Tied
Guides & Fundamentals8 min read

Where to Invest in AI Without Getting Vendor-Tied

Stop betting on AI vendors. Invest in three vendor-agnostic layers — tools, prompt workflows, and evaluation frameworks — that survive every model swap.

The AI Productivity Gap Your Readout Should Name
Opinions & Analysis10 min read

The AI Productivity Gap Your Readout Should Name

Engineering leaders lose credibility defending AI tooling with vendor ROI numbers. The honest quarterly readout names the gap and pitches the work to close it.

Your AI Skill Has an Eval Problem (Not an Adoption Problem)
Opinions & Analysis9 min read

Your AI Skill Has an Eval Problem (Not an Adoption Problem)

Engineering leaders blame stalled AI adoption on training or culture. The real bottleneck is evaluation discipline, the platform feature most rollouts skip.

Claude Code vs Codex vs Cursor (and Why It Doesn't Matter)
Opinions & Analysis9 min read

Claude Code vs Codex vs Cursor (and Why It Doesn't Matter)

The honest Claude Code, Codex, and Cursor comparison — and why your AI platform strategy matters far more than which coding agent you pick this quarter.

Beat Developer Burnout with AI-Powered Task Scoping
Problem & Solution6 min read

Beat Developer Burnout with AI-Powered Task Scoping

Stop drowning in task management. Learn how AI agents can automatically scope your work, research approaches, and maintain a stress-free pipeline.

Escape AI Platform Lock-In with a Portable CLI Toolkit
Tutorials & How-Tos4 min read

Escape AI Platform Lock-In with a Portable CLI Toolkit

Stop rebuilding AI tools for every platform. Build a portable CLI toolkit with centralized config that works with Claude Code, Cursor, or any agent.

Build Your Own Voice Assistant: Escape Alexa and Google
Tutorials & How-Tos7 min read

Build Your Own Voice Assistant: Escape Alexa and Google

Tired of creepy always-listening devices? Build your own private voice assistant with local wake words, any LLM, and custom Python tools on a Mac Mini.

Adding Eyes to Your AI Workflows with Gemini
Tutorials & How-Tos5 min read

Adding Eyes to Your AI Workflows with Gemini

Use Gemini's overlooked image detection features to find objects, extract precise bounding box coordinates, and build truly context-aware AI applications.

AI Content Pipelines: From Text Prompts to Images and Video
Tutorials & How-Tos4 min read

AI Content Pipelines: From Text Prompts to Images and Video

Build automated content pipelines using Claude Code skills and OpenAI APIs. Generate character images, convert to Discord emojis, and animate with Sora.

2025: The Year AI Development Grew Up
Guides & Fundamentals9 min read

2025: The Year AI Development Grew Up

From reasoning models to agentic CLIs, 2025 transformed how we work with AI. The timeline of our shift from model-centric to agent-centric development.