AI and Automation
Claude Code skills: when to use them and when they pile up as overhead
Skills load on demand and cost 75-150 tokens each in the listing. Use them for procedures over five lines. Skip them for project-wide rules.
Field notes
Field notes on AI-native products, design systems, web engineering, SaaS economics and modern infrastructure. New articles every week, sources included.
AI and Automation
AI and Automation
AI integration agency in 2026: what their pitch should actually contain
An AI integration agency pitch in 2026 is judged on three things: what gets deployed, how it gets measured, and how it gets handed back. Ten checks to listen for, five red flags to walk out on.
AI and Automation
12 mistakes we see teams make building their first multi-agent ops system
Multi-agent LLM systems fail in production at 41 to 86 percent. Most of the failures trace back to twelve specific decisions teams make in the first month. Here is what they look like and how to undo them.
AI and Automation
Claude integration partner: what to look for before you sign
A Claude integration partner ships production AI on Anthropic's stack. Here are the 8 checks, the 6 to 14 week timeline, and when you should not hire one.
AI and Automation
Multi-agent system design patterns that survive production
Multi-agent systems cost 15 times more than chat. 40% of pilots fail in 6 months. Five design patterns decide if your AI agent team ships or stalls.
AI and Automation
Claude Code hooks: PreToolUse, PostToolUse, and patterns that scale
How to wire Claude Code hooks for security gates, format-on-save, and team-wide guardrails without slowing the agent down.
AI and Automation
AI-native vs AI bolted-on: how to tell in 60 seconds
What separates AI-native products from retrofits: the architecture, the data flow, the team shape. Plus a 60-second test you can run on any product.
AI and Automation
Claude Code workflow patterns we use to ship production code daily
Seven Claude Code workflow patterns we run daily to ship production code: plan mode, CLAUDE.md, hooks, subagents, parallel sessions, slash commands, MCP.
AI and Automation
Build vs buy MCP server: when to write your own and when to buy
Self-hosting an MCP server costs $50k to $150k per integration per year. Buying solves 80% of cases, building wins the rest. Here is how to decide.
AI and Automation
How to build an MCP server for an existing Next.js SaaS
A pragmatic guide to shipping an MCP server on a running Next.js 16 SaaS: packages, routes, auth, session state, and the failure modes you will hit.
AI and Automation
What is MCP and why your SaaS needs an MCP server in 2026
MCP is the open protocol AI agents use to call your SaaS. Here is what it is, how it works, and when your product should ship a server in 2026.
AI and Automation
The most common errors when building apps with Claude
Seven mistakes we see repeatedly in production Claude apps. Some cost you tokens, some cost you money, one or two can cost you the company. Each one has a simple fix.