Just my thoughts about tech and AI, job market or in general about anything new i discover.
AI makes software output cheap, but it makes trustworthy engineering more valuable. In 2026–2027 expect heavy AI-driven code churn and market noise;
A personal walkthrough of adapting to Phoenix 1.8: DaisyUI, layout modules, async functions, custom modals, and JS tweaks for better UX.
In this article I share how I moved from random AI prompts to a simple flow for coding with agents: first make the agent understand the current code, then write a clear outline together, and only at the end let it change files. I explain how I use small guides stored in my project to teach the agent my style, how I reset context when chats get messy, and how I use the agent not just to write code but also to improve my own rules over time. It’s not about the “best” model, but about giving any model the right context and direction so it feels more like a sane helper and less like chaos.