Cubode’s Post

We're building an AI Agent in 30 days 😏 & documenting the whole process on Medium, YouTube, Instagram, Tiktok... 📹 If you are want to integrate Agents into your product but have no idea where to start, this is for you. All code included with step by step guides. Let's go!! #aiagents #ai #langchain #llamaindex #buildinpublic

🚀 Not even a month into the New Year, and it feels like product and design teams are really getting comfortable with the user experiences around AI. Here are three AI-enabled tools that are now part of my daily work: 🧙♂️ Superhuman AI: It's not about auto-written quick replies, which I try to avoid, but the nuanced features like timezone scheduling that make it a game-changer. 👨💻 GitHub Copilot Chat: So, you’ve definitely hear of Copilot, which is now in GA. But the recently added chat feature in IDEs like VSCode is a revelation for hobbyist coders like me. It explains concepts, reads code, and, announced at GitHub Universe in October, will begin integrating third-party dev tools soon. 🏄♂️ Perplexity.ai: This tool is combines search and LLM-chat in a unique and differentiated experience. I find that getting the answer I need comes quicker than either search or ChatGPT. Still, it raises questions about the business impact on publishers. What AI tools are you using on the daily? Let’s discuss. #AI #ProductivityTools https://lnkd.in/etsf7zgj

As folks are racing to build their own LLM solutions, a common question I've been asked is— what are the tools out there to streamline this process? So I wrote a blog 🚀 - where I demonstrate how to "Build A Custom AI Based ChatBot Using Langchain, Weaviate, and Streamlit"! 🤖📚 Whether you're a developer, AI enthusiast, or just curious about the magic behind AI-powered conversations, this blog has something for you. 🔍 Here's a sneak peek into what's covered: 1️⃣ LLMs with Retrieval Augmented Generation: Learn how to leverage the power of ChatGPT, GPT-4, or your favorite Language Model to create dynamic and context-aware responses. 2️⃣ Vectorization and Databases with Weaviate: Uncover the secret sauce behind extracting pertinent information from a sea of data. See how Weaviate comes to the rescue with its vectorization capabilities, making information retrieval smarter and more efficient. 3️⃣ Prompt Engineering/Chaining via Langchain: Dive deep into the technique of prompt engineering and chaining to guide your AI companion's conversational flow. Langchain's toolkit is your ally in shaping meaningful interactions. 4️⃣ Sleek Frontend and UI Powered by Streamlit: Learn how to wrap your AI chatbot in an attractive and user-friendly interface using Streamlit. Impress your users with a polished, interactive experience. 🔗 Link to the full blog post: https://lnkd.in/gR4qJ2vJ Let's unleash the potential of AI together! 🌟🤝 GitHub code in the comments! #AIChatbots #GenerativeAI #CustomSolutions #TechTutorial #StreamlitMagic #Weaviate #Langchain

Last week I mentioned Devin, and while we wait to try it out, I decided to tell you how I use AI during development, and well, maybe get some new tools/suggestions from you. If it's about me, almost every day I use: 👉 Copilot, of course! It's like an extra pair of hands. It's decreasing the need to think about syntax and sometimes even create whole functionalities for me. With it, I can completely focus on building, without getting tripped up over the basics of syntax. 👉 Then, there's ChatGPT – where do I even start? It helps me to generate new ideas to test, prepare work plans, adjust messages, or lend a hand with codes that aren't within my strict specialization (like tricky MongoDB query optimization or improvements or code explanations). And, of course, once I need to create content (and we all know it's not a dev's favourite thing to do) it helps me to make it efficiently by polishing the technical language I use :) Of course, I also use AI to automate the processes. Have you seen what AI can do with Storybook code? There's a super cool guide on how to roll your own Storybook GPT (check it out here - https://lnkd.in/dwcvBvSR). And that just makes my work smoother. And, I can admit - more fun! I also like experimenting with new tools. Right now, I'm checking v0.dev, to see if they can do some initial work for me. I'd love to hear which tools you're using! Catch you in the comments! #ai #development #devtips

🚀 Excited to share my latest project: An AI-Powered Real Estate Data Analysis Platform! 🏡💡 Over the past few months, I've been working on a revolutionary backend application designed to transform how we analyze and interpret real estate data. Built on the robust Bun.js framework and powered by OpenAI's ChatGPT 4.0, this platform leverages the power of AI to generate insightful queries and provide data-driven investment suggestions based on user input. Key Features: 🔍 Dynamic Queries: Our AI dynamically generates queries based on user input, allowing for a highly personalized data analysis experience. Whether you're looking for high rental yield properties or the best investments in a specific city, this platform has got you covered. 📊 Data Summarization: The platform excels at summarizing large datasets, distilling complex real estate information into easy-to-understand summaries that highlight the most critical aspects of the data. 🔎 Insightful Identification: Our AI identifies key trends and patterns within the data, helping users make informed decisions by highlighting significant insights and potential opportunities. 📈 Accurate Projections: By leveraging advanced algorithms, the platform can project future trends and market movements, providing users with valuable foresight into potential real estate developments. 💡 Smart Suggestions: Based on the data and user preferences, the platform generates smart investment suggestions, guiding users towards the best opportunities in the market. Why Bun.js and ChatGPT 4.0? Efficiency: Bun.js provides a lightweight and highly efficient backend framework, ensuring smooth and fast performance. Intelligence: ChatGPT 4.0 brings advanced AI capabilities, enabling us to deliver highly accurate and relevant insights. This project is a testament to the incredible potential of AI in transforming industries. I'm thrilled to see how this platform will empower real estate professionals and investors to make smarter, data-driven decisions. Stay tuned for more updates and insights as we continue to enhance the platform's capabilities! Best, Brandon "Metavibez" Jackson #AI #RealEstate #DataAnalysis #MachineLearning #BunJS #ChatGPT #Innovation #Technology #Investment #Startup #TechTrends #SmartInvesting

🚀 Build in Public - Day 76 Update 🚀 Today, I've got some exciting news to share about our progress on TryHumanize and an intriguing development with Claude 2: 📝 Article Outlining: We're diving deep into building the infrastructure for our upcoming Article Outlining feature. Once the user interface (UI) is complete, we'll seamlessly integrate the backend. I'm genuinely thrilled about this addition because it's all about putting you in the driver's seat when it comes to content generation. With Article Outlining, you'll have the power to tweak and mold the article outline as per your vision. This level of control will enable you to craft articles precisely to your liking, optimizing the content generation process and conserving tokens on both ChatGPT and Claude 2. 🚀 Claude 2: Here's a sneak peek into a fascinating development - I now have access to Claude 2, and let me tell you, it's a game-changer! I'm currently putting it through its paces, generating blogs and exploring its capabilities. Claude 2 is blazingly fast, and the quality of its output is on par with the legendary GPT-4. While the API integration will take some time, I can already foresee the tremendous potential it holds. These developments are all about enhancing your content creation experience, giving you more control, and providing access to cutting-edge AI technologies. I can't wait to see how these additions empower you to create exceptional content effortlessly. Keep your feedback coming, and stay tuned for more exciting updates! #BuildInPublic #TryHumanize #ContentCreation #ArticleOutlining #Claude2 #AI #Innovation

Most GenAI apps are “cheating” to answer questions. Here’s how and why companies intentionally help them do it: First, take a common knowledge topic and ask Gemini, ChatGPT, or whatever large language model (LLM) you like to answer it. Something like “who won the 2013-14 Super Bowl” should work well. The model will be able to answer it, no problem. Next, ask about something personal, like: "Where was Michael Verkruyse's last vacation." It won't know because it doesn't have that data in its training context. But what if we “cheated” by providing that context? For example: "Where was Michael Verkruyse's last vacation. Context: Michael has a 4-year-old and a 1-year-old, so 'vacation' might not feel like the right word anymore, but his last trip was to Seabrook, WA with his family." Now, the model can answer correctly! It's not exactly "magical"—we essentially handed it the answer. But that's precisely what companies are doing to enhance LLMs—they're supplying the additional context needed for accurate responses. However, with companies, there's a vast amount of data that could serve as potential context, and most of it is irrelevant for a specific question. Using all possible data can lead to extra costs, slower application speeds, and less accurate answers. To find the right data for each question, most GenAI applications use a process called Retrieval Augmented Generation (RAG). RAG involves several steps that can be finely tuned for specific use cases, but fundamentally, it includes: - Preprocessing - Chunking - Embedding - Storage - Retrieval - Generation These help give the model the context it needs to “cheat” and provide the user with the best answer. In the spirit of cheat sheets, I’ve included a diagram that breaks the steps of RAG down further, so you have the context needed to answer questions about it if you ever get asked yourself! Feel free to share if you find it helpful! 

Hi there, Network! 👩💼 I am really enjoying LangChain at the moment, so today I bring you my latest article on LangChain v0.2 {Agents} and {Tools}! I want to help you build powerful assistants that use ChatGPT. Imagine having a model that can understand user inputs and call other functions, including your own models! Check it out here: https://lnkd.in/dvFgB9kv Let's make our models more accessible and useful! 🔗 #LangChain #ChatGPT #MachineLearning #AI #LLM #DataScience #RealEstate

StackOverflow is fighting back! Here is everything you need to know about the release of OverflowAI, their new generative AI tool. StackOverflow traffic has been declining for several years, but after the release of ChatGPT, things took a turn for the worse: Posting activity on the site fell the equivalent of five years in just six months! People stopped visiting the site and asking questions. Many thought this was a slow and painful death, and there was no coming back. Today, their CEO announced OverflowAI. If you can't beat them, join them, right? But what's OverflowAI, and is this enough for StackOverflow to become relevant again? Here is StackOverflow's statement from seven months ago: "All use of generative AI (e.g., ChatGPT and other LLMs) is banned when posting content on Stack Overflow. This includes asking the question to an AI generator then copy-pasting its output as well as using an AI generator to reword your answers." OverflowAI is not a tool but a group of different generative-AI initiatives. StackOverflow went from banning AI to embracing it in over six months. Some of the features they announced will be open to everyone: First, they will replace their search functionality with one powered by generative AI. You can ask a question, and the site will search and summarize existing answers for you. To power this feature, they use a vector database and embeddings of their 58 million questions and answers. It's a smart move, and it should improve the quality of the solutions we get. Answers will cite sources, so authors get the credit, and users can always trace back the results. Second, the assistant will help you ask a question if you don't find an answer. This should help people keep a consistent format and include as much relevant information as necessary. The third interesting feature is a Visual Studio plugin to ask questions from your IDE. You can select a snippet of code, fire the plugin, ask a question, and get an answer without opening a browser. I'm having difficulty seeing how a plugin that answers questions will fit next to Copilot. Will they complement each other? I'm glad StackOverflow is fighting back. They promised the new search functionality by August. But is this enough to keep them alive? Do you think they have a fighting chance against ChatGPT and Copilot?

#ai buzzword update: #RAG, #AGI. Almost everyone has heard about ChatGPT by now. Depending on where you are on the adaptation curve you have might have talked about it (a LLM based Chatbot), tried it a few times, uses it regularly or even tinkered with it for business use cases. The number of ai-solutions in production are still very limited however, but MS Fabric will probably change that by making it much easier to go from test to production (if pricing becomes transparent). However things move quickly in the generative ai world and I wanted to draw your attention to: RAG, or "Retrieval Augmented Generation", which is the enhancement of the chatbot you know. You provide (augment) it with context in the form of data and instructions to guide the responses it should give. This eliminates much of the hallucination (making up stuff) that has been talked about in the media. This is possible right now using programming frameworks like #langchain, but both Bing chat and OpenAI in the paid version offers it directly without coding. AGI, or "Artificial General Intelligence" is on the other hand not here (yet). Using RAG and multimodal (images, texts, code etc.) models comes close to simulating it, and then you can ask yourself if that makes a difference if its real. Things might change soon though, and that is where the article below comes in. Apparently OpenAI has succeeded in implementing some of the RAG ideas into the core ai model and thereby sparked the first steps towards true AGI. This will be the conversation for Christmas next year ;) https://lnkd.in/dF-yjPUV