Ask HN: How are you using LLMs in your personal projects today?
I started working with AI / LLMs recently, and now I'm learning about the ecosystem. The possibilities seems unlimited and I know many use cases for big companies. However, I'm curious to learn about what people are hacking with LLMs on a personal side.
What are you building? What are you solving with LLMs? How do you run them? How do you integrate them in your app / service?
Comments
I am mostly running local models using Ollama on a 32G on chip memory Mac Mini. This morning I am experimenting with function calling. Yesterday was a RAG app with a simple web UI.
Historically I have worked in NLP, and LLMs have made many older techniques obsolete, and that has been fun.
I have also been using OpenAI APIs for about 2 years and now I also experiment with Anthropic’s Claude 2 and Google Bard. Nothing important, I just want to have a solid intuition on what different commercial models can do.