Hey guys! I just sat down with @Patrick Chouinard for one of the coolest deep dives we've done in the community. Patrick has become our community's go-to AI expert and one of the most helpful resources we have. He's quietly been building something most teams pay a vendor $50k a year for. He's turning every community call we've ever recorded (two and a half years of two to three hour conversations) into a queryable "community brain." You'll soon be able to ask it anything that's been discussed and get a real answer with citations. Here's the part that broke my brain. Standard RAG completely falls apart on transcripts. The question gets asked at minute 6, the conversation drifts, and the real answer shows up at minute 41. A normal chunker has no idea those two moments belong to the same idea. Patrick solved it by adding an LLM analysis layer BEFORE chunking that restructures the transcript into self-contained units of knowledge. You can watch the full breakdown above! Here's everything we covered: ✅ Why traditional chunk-and-embed fails on non-linear data ✅ The LLM analysis layer that turns raw transcripts into RAG-ready knowledge ✅ How Patrick picks the cheapest model that's still smart enough for each step (Kimi K2.5 for restructuring, Sonnet 4.6 for the signal extraction) ✅ Why he chose LanceDB over Pinecone (and when you'd flip that decision) ✅ Running the whole thing locally with Ollama, Open Web UI, Gemma 4 4B, and gpt-oss:20B for more complex retrieval ✅ Using Claude Code to build a custom chunker instead of fighting with a library ✅ Real cost math. About 40 cents per two hour episode and under $100 to process the entire archive The wildest part is the price. Once the embeddings are built, querying is free forever because everything runs on your own machine. No SaaS, no per-token cost, no IT review. Patrick is also planning to open source the full pipeline once a few rough edges are ironed out. So if you've been wanting to build something like this for your own team, agency, or client, you'll have a working blueprint to start from.

Unfortunately for Cursor, Grok was built to be so truthful that it couldn't lie to protect it's secret identity 😅 On a more serious note, this model is an absolute beast! Usually when working with larger 1m context window models, you tend to give up speed so that you can cram more info into the context window. So far, code-supernova-1-million has been insanely fast and has been able to handle medium sized coding tasks very well! Now that Grok is cheap, fast, smart, and has a large context window, I could easily see this new model to go to model for every day development tasks. If you get a chance to try out the model, please let me know what you think!

Guys! I'm so excited! 15 more hours till shipkit.ai officially goes live! I'm doing a final round of tests just to make sure that every template works perfect for you! So far, everything is going smooth as butter! If you're thinking about building your own AI application for your own start or to start AI freelancing, shipkit is perfect for you. You are going to be blown away with how easy it is to setup, build, and deploy your own AI application. Over 150 developers have signed up for the prelaunch to jump start their AI Development journey! We'd love to have you as a part of the shipkit fam! Make sure you don't miss out on the $100 prelaunch discount that goes away as soon as shipkit launches here in 15 hours! You can grab shipkit here: 👉 https://shipkit.ai

I am looking for a mentor to walk through my project with me, review design, troubleshoot etc. I'm happy to do the work but need guidance and advice etc. If you have an interest in teaching others or are interested in being a consultant please contact me, is live to go over the project and get some feedback.

🚀 Big Announcement In just 10 days (Friday, September 27) we’re officially launching something I’ve been waiting to share with you for a long time. It’s called Shipkit.ai. 👉 You can pre-order today and lock in $100 off before launch 💡 Why I built Shipkit.ai Over the past few years I’ve taught thousands of developers through my YouTube channel (75,000+ subscribers) and inside my community (nearly 10,000 members). I kept seeing the same roadblocks stop good ideas from ever shipping: - Which stack should I choose? - How do I connect all the pieces? - I built a demo on my laptop… but how do I deploy a real product people can actually use? Advice helps, but it isn’t enough. Developers don’t just need tips — they need working projects. 🛠️ What Shipkit.ai gives you Shipkit.ai gives you everything you need to go from idea → production-ready AI app. You’ll start with working templates for the most common AI applications: - AI chat - RAG - ADK workflows Shipkit walks you through each one step by step so you know exactly how to customize and launch it as your own. No guesswork. No dead ends. Just shipping. The course inside Shipkit shows you every stage of the journey: - Idea → Development plan - Plan → Working code - Code → Deployed application in production And at every step, you also get prompt templates that show the AI how to take action for you. These help you move faster and avoid getting stuck so you can focus on shipping real applications instead of wrestling with the process. ✅ Proof it works Shipkit.ai was built with Shipkit. To prove it, I used the RAG template inside Shipkit to build this very platform. That means you’re not just getting theory — you’re learning from the exact system that produced Shipkit.ai itself. The same step-by-step process that takes you from: - Idea → Plan - Plan → Code - Code → Production