Hello everyone, hope you’re doing well! I’m building a hybrid RAG pipeline for corporate environments—combining lexical retrieval with sequential semantic filtering, and now adding a graph‑view layer to further enrich the context. At this final stage, I’ve encountered some interesting challenges, such as: - Balancing precision and performance when crossing textual searches with vector searches; - Correctly modeling graph edges, ensuring we only link truly relevant elements; - Normalizing the input format so downstream components (semantic filter, graph retriever) communicate smoothly. The goal is to bring together the best of both worlds—classic text search techniques and powerful vector‑similarity filters—and ultimately design a knowledge graph that drives corporate insights (relationship flows, entity‑topic maps, etc.). If you have any suggestions, questions, or want to share experiences about RAG, knowledge graphs, or hybrid pipelines, feel free to comment here or reach out on LinkedIn. 👉 Let’s connecthttps://www.linkedin.com/in/rtcoutinho Cheers!

Everyone’s building wrappers.We’re building a new reality. I’m starting an ai powered Social network — imagine X or Instagram, but where the entire feed is 100% AI-generated.Memes, political chaos, cursed humor, strange beauty — all created inside the app, powered by prompt. Not just tools. Not just text.This is a social network built by and for the AI-native generation. ⚠️ Yes — it will be hard.But no one said rewriting the internet would be easy. Think early Apple. Think the original web.We’re not polishing UIs — we’re shaping a new culture. We’re training our own AI models. We’re not optimizing ads — we’re optimizing expression. 🧠 I’m looking for: - AI devs who love open-source (SDXL, LoRA, finetuning, etc.) - Fast builders who can prototype anything - Chaos designers who understand weird UX - People with opinions on what the future of social should look like 💡 Even if you don’t want to code — you can: - Drop design feedback - Suggest how “The Algorithm” should behave - Imagine the features you’ve always wanted - Help shape the vibe No job titles. No gatekeeping. Just signal and fire. please email me to help out vilhelmholmqvist97@gmail.com

🗞️ BREAKING NEWS : ▪️ OpenAI CFO Sarah Friar recently revealed that the company is set to launch its third AI agent. ▪️ The new agent is called “A-SWE”, short for Agentic Software Engineer. A-SWE will be capable of performing tasks typically handled by human software engineers. ▪️ In addition to coding, it will also manage tasks such as quality assurance, bug testing, and bug bashing. ▪️ This marks a significant advancement in AI’s ability to contribute across the full software development lifecycle. — — — — - ➡️ I am curious to learn how these AI agents will actually help with bug test case preparation, identifying corner cases, handling out-of-bound scenarios, and even fixing bugs. ➡️ Bug fixing is rarely straightforward — it often requires understanding which module is causing the issue and where exactly the fix needs to be applied. I am especially interested in how AI can perform root cause analysis, since that typically involves deep investigation across multiple layers of code and logic. 🖊️ ✴️ IMO, AI agents can help with debugging by creating tests, finding edge cases, suggesting fixes, and tracking down problems. 😊 However, the final decision still needs a human engineer’s logic, intuition, and expertise. 😊 — — — — PS : Stop Video at 11:28 and listen.

Y-Combinator's recent Requests for Startups : https://www.ycombinator.com/rfs

https://youtu.be/M80nBMmHywc?si=7xYhM86072IXCM1T watch the video and share🌟❤️🇺🇸 I’m at the Disaster EXPO IN MIAMI. I need your help getting FEMA to see the video so we can make it happen. Any promotional help is greatly appreciated❤️🇺🇸🙏🏼🌟 🚨 FEMA, It’s Time to Revolutionize Disaster Preparedness! 🚨 🌪️🔥🌊 Rescue Connect: The Life-Saving App That Puts Citizens First! 🇺🇸 Imagine a world where every American is prepared BEFORE disaster strikes —where supply chains, emergen