What "real data" are you searching for through Perplexity for the AI you use? The RAG functionality of Perplexity isn't as reproducible as a full RAG process that I'd control, but it's super flexible for public internet data. What do you use Perplexity for? What would you want to automate around it?

After two years of building with GenAI, here’s what I wish I’d had from day one... This has been a long time in the making, and I’m excited to finally pull back the curtain on what we’ve been building behind the scenes: The GenAI Launchpad — officially launched on Product Hunt today! 🎉 For the past two years, my team and I at Datalumina have been deeply involved in the world of AI, building solutions with large language models (LLMs) for clients across industries. Each project taught us so much about what it takes to bring AI to life in practical, high-impact ways. But there was one recurring challenge... We spent way too much time setting up project structures, handling integrations, and putting out fires in the infrastructure — leaving less time for the real AI work, the work that brings ideas to life. Not only was setup eating into our time, but we also found that the agent frameworks on the market were just too optimistic. Real-world use cases are more complex and demand reliability and precision that many frameworks simply can’t deliver. So, we got to work! 👷🏼‍♂️ And after two years of trial and error, working with every system and structure you can imagine, we built our own solution. The GenAI Launchpad is the result of our journey — a project repository that streamlines everything from initial setup to deployment, ready to handle the demands of production at scale. And the time savings? ⏳ We’ve calculated that it saves us over 50 hours per project on average, so we can dive right into the creative work that actually advances AI. Today, we’re launching the GenAI Launchpad to share that time-saving power with you — our community of fellow AI enthusiasts and builders. This is more than just a repository; it’s a battle-tested, engineer-approved blueprint that I wish I’d had when we started. It’s here to help you skip the headaches, bypass the boilerplate, and focus on what matters: building innovative AI solutions for real-world problems. If you’ve ever spent weeks fighting project setup, only to finally reach the real work, then you’ll understand why I’m so excited to share this.

People or bots (IDK) from different accounts keep messaging me about dropshipping and e-commerce. I dont know if they want to sell a course or what… Are any of you receiving these kind of messages from different useres?

I was talking to someone in AI who is somewhat technical, but they didn't understand Vector Embeddings. I tried to explain them, but felt my explanation was lacking, so I'm going to try again. Let me explain Vector Embeddings in the simplest way I know how. On a scale of 1 to 10, how happy are each of these phrases in the screenshot? Congratulations, you just created a one-dimensional Vector Embedding of those phrases. You can plot this on a chart and look for similar amount of happiness. You can find happy phrases or less happy phrases just by looking for closeness to a certain number. That's simple, right? But, not too helpful. Let's add another ranking something around what the phrases are about. From 1 to 10, how "about clothes" is the phrase? Look at you, you created a two-dimensional Vector Embedding now. You can also chart these so you can visually look for nearby phrases for similarity. This helps you find the happiest ones about clothes! This could also be useful for looking at a bunch of reviews and looking for a relationship between happiness and some feature of the product, like the comfort of it, so see if people like the feel of it or not. Now, to get a bit more technical: We could go on and on here, adding tons of scales (also called features or dimensions). We can show it in 3d, but depth can be tough to see on a screen. We can add colors, showing green as positive and red as more negative we. Or we could change the sizes of the dots, making higher numbers bigger. That's just six scales. And you can probably do a lot with just six scales. So, that's what a Vector Embedding is. There are all kinds of uses for it with computers, but they do what you've already done here, help you find similar phrases. But, you want to know something cool? OpenAI's text-embedding-3-small has 1536 of these scales. Wow! And there are a lot of different ways to determine these scales, different models use different ways. People will try the different versions and see which ones do better for whatever they're doing. Instead of drawing these on a chart, some people store them in a Vector Database, which keeps and finds this data really well. Instead of looking for close ones on a chart, which doesn't really work with more than six scales, they use a math equation to find nearby phrases. One of those equations is called Cosine (pronounced Ko-Sign) Similarity.

When building RAG systems, embeddings + text search is better than embeddings on their own Combining these techniques together is called Hybrid Search. In this tutorial, I walk you through building a hybrid search pipeline using just PostgreSQL. On top of that, we'll also look at how we can further improve the accuracy using Reranking. Curious how a hybrid system works and how it can optimize your search accuracy for AI applications? Watch the video here P.S. I was supposed to record this from my new office, but, well… life had other plans. So instead, enjoy this kitchen table special — different setup, same Dave vibes. I think it turned out pretty alright!