During my time at Berkeley, I had explored how we could use crypto and live-streaming to address the problem of echo chambers on social media. This project is another shot at a that problem. The quest for people’s attention makes media businesses pick sides and often leads to biased narratives that cater to what their consumers want to hear. I always wondered if we could build something that provides us with just the facts, without opinions or bias. Turns out we could potentially hope to do that using Large Language Models (LLMs). That’s why I wanted to explore building a RAG-based LLM application to fetch and summarize relevant news in concise format.

There are a lot of good resources out there on building LLM applications. So, to avoid repeating the same in this post, I’ll just focus on what I tried to build and my experiences. I will update this post shortly. In the meanwhile, you can find the code here: GitHub.

Profile Pic