ClemDev2000

In this post, we’ll build a lightweight code agent CLI, similar to Claude Code.

By using a vector database, we can bypass the finite context limitation of current LLMs and give the model only the context relevant to the task in hand.

This blog post gives more details on the science of training large autoregressive language models that have recently been successful worldwide (ChatGPT, GPT-4). Some thoughts on the similarities between the biological brain and artificial neural networks are also presented.