In this livestream, I’m walking through what “building AI apps” looks like in 2026—because the workflow has changed a lot since “vibe coding” first became a thing. Back then it was basically: prompt, accept whatever comes out, and hope for the best. Now it’s much more structured: agentic engineering (or honestly, I just call it AI coding) is becoming the default for professionals, with more oversight, testing, and reliability so you can actually ship production-quality software without compromising quality. I also break down the tool landscape from no-code all the way to pro dev, and why open source is a huge unlock—not just for cost, but for privacy and new use cases. I show an example internal tool my team built: a slide generator that creates HTML-based slides, does research, and cites sources to reduce hallucinations. Then I zoom out into the kinds of AI apps you can build beyond chatbots: automation agents, document processing, content creation, analytics, and creative tools. My favorite category is “hyperspecific AI apps”—software that’s custom to your exact workflow (like bookkeeping for a messy international business setup), often cheap or free after setup, and sometimes unlocks things you simply couldn’t do before.