Tiny Software Factories: Agentic Engineering DONE WRONG
What's up Engineers, here's a SNEAK PEEK of the upcoming Phase 3 Agentic Engineering Course: Tiny Software Factories Phase 3 PRESALE IS LIVE (What if your agents never stopped?) https://agenticengineer.com/infinite-agentic-engineering?y=1v7ldJiyH-Y Today you'll learn agentic engineering DONE WRONG. On purpose. 🤨 You know this... You and I are the bottleneck. Every time you sit in the terminal prompting one agent back and forth, YOU are the system. Eighty percent or more of the work we're doing right now can run without us with the right system in place. In loop, you prompt. Near loop, you respond. This is blitzscaling for agentic engineering. It's not meditation, it's Xanax. It's not working out, it's a GLP-1. A useful local maximum for getting results at all costs. We stop worrying about getting things right and start getting things SHIPPED. 🏭 WHAT IS A TINY SOFTWARE FACTORY? A tiny software factory is a software factory built to solve one specific problem where a proactive agent works around the clock in a durable environment with you near the loop instead of in it. One problem. One proactive agent. Around the clock. Durable. Human near the loop, not in it. ⚠️ AGENTIC ENGINEERING DONE WRONG Tiny Software Factories is a crash course in agent anti-patterns that will NOT scale, and that's the point. The shortcuts that make a factory useful today: one agent, cron in the server, a JSON config, agent-written pages, SQLite on one box, and bash in your codebase. Each one gets you from zero to value fast. Each one breaks somewhere, and you'll know exactly where. 🚀 In this lesson, we open the empty factory: a fully built template with every feature you need and no specialization yet. Then we stack features from the feature map, ranked by value and effort: 🤖 The agent on your right: a durable agent, always on, always one message away 📋 Lists on your left: every product is a list of something. What's in your list? 🎯 Focus in the center: a generative UI where your agent builds, tweaks, and tunes pages for you 🔁 Agent access parity: everything you can do, your agent can do. Every UI control has a matching agent tool 🧠 Core four control: context, model, prompt, tools. Auto compaction, tiered model selection, a prompt engineered system prompt, every tool in view 🧩 Extensible harness: your agent is a composition of capability slices (factory, coding, pages, image generation, factory controls) 🎨 Customizable DX and UX: themes, nav bars, hotkeys, animations. Liquid software, built to be changed ⏰ Cron prompts: prompts on a schedule. Close the browser and your agent keeps working 💡 Watch the agent modify ITSELF: one prompt disables all of its coding tools and raises its own thinking level. Reflection for AI agents. Powerful, and dangerous if you don't control it. 🛠️ INITIALIZE, CONFIGURE, SPECIALIZE With one meta skill and one command, you mount a new factory next to your codebase. Point it at your repo, drop it in a sandbox, and go from zero to a working agentic system in minutes. Every prompt, tool, hotkey, extension, and theme is accessible to you and your in-loop agent, so you build the system that builds the system. 🔥 TSF vs IA3 Tiny Software Factories is a highly compressed, lossy version of the Infinite Agentic Engineering core lessons. TSF lives near the loop: fast, rapidly deployable, intentionally imperfect. IA3 goes out the loop, where you build fully autonomous systems that operate for you without you. Know which one you're building, and why. ✅ Key takeaways: In Loop vs Near Loop: Stop babysitting AI coding agents. Give feedback, make decisions, and approve work your agent has already done. Data Flywheel: Every signal you give your near loop agent makes the factory better. Proactive Agents: Durable agents plus a trigger inside a sandbox. It doesn't wait for you to work. It works for you, then waits. Agent Cron Jobs: A name, a schedule, a prompt. The fastest path to repeat work running while you're away. Right Model for the Job: Don't fixate on one model. Whether you route through OpenRouter or run a pi agent harness, pick the right performance, cost, speed tradeoff. Join IndyDevDan as we build tiny software factories that work around the clock, so you can get things done at the speed of agents. Stay focused and keep building. IndyDevDan #agenticengineering #softwarefactory #aiagents
