Cursor Review After Three Months of Daily Coding
A pragmatic Cursor review for indie builders coding 4 plus hours a day. Pro tier math, agent reliability, and when the $60 Pro Plus is worth it.
How to Split Your AI Agent Across Two Model Tiers
A working pattern for cutting AI agent costs 70-80 percent. Route the cheap calls to a small LLM, send synthesis to a premium one. With code.
Build Memory Architecture for Long Running AI Agents
How to build memory architecture for long running AI agents that survives sessions, model swaps, and changing user preferences in production.
Build a Secure GitHub PR Agent With Claude in a Weekend
Build a GitHub PR auto fix agent that ships actual fix PRs instead of comments. Step by step with Claude, GitHub Actions, and security patterns.
GPT-5.5 vs Claude Sonnet 4.6 for Agents and Code
GPT-5.5 vs Claude Sonnet 4.6 compared on agentic benchmarks, coding, hallucinations, pricing, and speed. Specific picks for builders inside.
Build a 14-Skill AI Job Search Agent With Claude Code
Build an AI job search agent with Claude Code using the 14-skill pattern, Playwright portal scanning, and per-listing CV adaptation. Real builds inside.
Owl Alpha Review When Free Comes With a Speed Tax
Owl Alpha review covering the free 1M-token context window, 12 tps speed tax, 4.7 percent tool error rate, and who should really use it.
How To Build AI Voice Agents Clients Pay $5k For
How to build production AI voice agents on Vapi that paying clients keep. The 4 skills, the sub-500ms stack, turn-taking config, and the price math.
How I Built a Daily Brief AI Agent With a Receipt Printer
A father built a daily brief AI agent that prints to a thermal receipt printer with one button per kid. Here’s how to replicate the whole setup.
Run Claude Code 24/7 With an Approval Queue Pattern
How to wire Claude Code into a 24/7 approval queue with a Telegram bot, file-locked state, and Stripe fulfillment, based on a real production agent build.
