This one is the build that goes with yesterday's theory article.

You will write a Python script that holds a conversation with an LLM, watch the messages list grow with every exchange, and see for yourself why the model seems to remember things. Step by step with everything explained.