I’ve put together a video to show how the JetBrains AI assistant works. This tool integrates with the IDE, letting you describe what you need in a chat-like interface. From there, it generates the code, including files and fields, all in one go. About two years ago, back in early 2022, I experimented with the GitHub Copilot Technical Preview, which relied on prompts in comments to suggest code snippets. Since then, I’ve been using it in my daily work, and it’s been a great tool to have. But it feels stale now and very limiting.
The JetBrains approach feels different as you’re directly telling it what to build, and it sets up the whole system. I had an unfinished project, just a rough starting point, and used this assistant to complete it in a couple of hours. This old project had a database abstraction layer I’d built to operate on the data, but I’d only implemented it in one or two places as examples. By showing the AI assistant the abstraction and how I used it in those initial examples, it was able to go and modify the rest of the entire codebase following the same pattern I had established.
This was the big jump I noticed from GitHub Copilot. JetBrains AI Assistant can modify entire files from a single prompt, understanding the broader codebase thanks to its built-in vector database used for RAG (Retrieval-Augmented Generation). This really feels like what the future of programming looks like, which I think is way less boilerplate and more fun actually seeing your project progressing fast.
For more thoughts on how AI is changing development, check out my article on how AI is changing art and programming. It’s an interesting shift in coding tools, so take a look if you’re curious about how it functions. This development approach aligns well with modern DevOps practices where automation and efficiency are key. The ability to quickly generate and modify code is particularly valuable when working with containerized applications or when you need to rapidly prototype new features.