The author shares their journey of setting up a locally hosted large language model using tools like DeepSeek R1, Deno, Jupyter Notebooks, and LangChain.js. They explore the compatibility of TypeScript/JavaScript with AI technology, showcasing the efficiency and ease of experimenting with AI using Deno as a development tool. The instructions include setting up the local model, creating a Jupyter notebook, generating AI “chains” using LangChain.js, and validating the model’s output with a JSON schema. The author concludes that Deno and Jupyter provide a smooth environment for AI experiments, highlighting the benefits of low-friction development and thorough documentation.
https://deno.com/blog/the-dino-llama-and-whale