Summary: Turning MicroGPT into a more educational, configurable project and finding that modifying and experimenting with it was the best way to understand how LLMs work.
Summary: Upgrading a long-neglected Django project turned out to be far smoother than expected, reinforcing the Django project's strong commitment to backward compatibility.
Summary: Hands-on account of testing GitHub Copilot Pro, Claude 4.5, and GitHub's new review tools by delegating a real Django feature implementation to the AI, with minimal manual coding.
Summary: LLMs can't reliably do math or counting on their own, but tool calling lets them delegate those tasks to external code or APIs, making them accurate without needing innate calculation skills.
Summary: Why Go is a compelling choice for building CLI tooling, argumented with an example project.