So everyone's doing their 2026 goal-setting thing, right? I went a different route. I asked Claude: "What am I going to regret 12 months from now?" Here's the exact prompt I used: "It's December 2026. I'm looking back at this year. Analyze all our chats. Based on my current trajectory, what will I regret NOT doing? What will I wish I'd said no to? Be specific and brutally honest." And honestly? The feedback was... uncomfortable. 70% was spot-on enough that I had to sit with it 20% made me want to argue (which probably means it's hitting a nerve) 10% was off because Claude was missing some context The parts that really got me: "You're building everyone else's systems. When do you build the thing that's unmistakably YOURS?" "You're too available. That doesn't build wealth or freedom." "Stop saying yes to generic AI training workshops. You're positioning yourself as a commodity." Ouch. But also... true. Why I think this beats regular goal-setting Goals ask: "What do I want?" Regret asks: "What will I actually wish I'd done?" That second question? It cuts straight through all the BS we tell ourselves. It shows you what you're REALLY doing vs. what you think you're working toward. Try it yourself? → Use whatever AI you chat with most (Claude, ChatGPT, whatever) → If it has memory turned on, just paste the prompt → If not, give it some context first (screenshots of your calendar, recent project notes, whatever shows what you're actually up to) Not everything it says will be right. But the stuff that makes you defensive? That's the good stuff. Anyone else brave enough to try this? What did your AI roast you about? 👇

Today I reached my goal. Thirty days of focused learning on RAG. And actually, if I look back, it has been even longer.November was already a full month of Claude Code learning, and December became my second one in RAG. This journey has changed the way I think about RAG systems I am incredibly grateful to everyone who supported me, answered questions, shared insights, recorded Loom videos or simply encouraged me to keep going. You made this learning journey possible. And of course, the road does not end here. Now I want to go deeper into: • Claude Skills • Claude Code • MCP • GoHighLevel • and of course… n8n But before I move on, I want to look back at what I actually learned in RAG. • Chunking • Embeddings • Reranking • Metadata • Hybrid RAG • Graph RAG And the tools I worked with • Pinecone • Supabase • Cohere • Google File Search From zero to building real pipelines, assistants and automation flows. Sometimes frustrating. Very often exciting. And always worth it. RAG will continue to be part of my path, but now it becomes one piece of a bigger picture. Thank you all for being part of this journey.