From the course: Azure Infrastructure as Code with ARM, Bicep, Terraform, and Pulumi

Impact of AI on IaC tooling - Azure Tutorial

From the course: Azure Infrastructure as Code with ARM, Bicep, Terraform, and Pulumi

Start my 1-month free trial Buy for my team

Impact of AI on IaC tooling

“

- [Instructor] AI has impacted the world in so many ways. It's hard to gauge just how much it will impact as we go forward, but where we are today is that we have large language models that are now delivered through many different services and that we can use to help us understand, describe, and even generate code. For the developer, working with cloud infrastructure, this means that we get the benefits of things like IntelliSense on steroids. We get suggestions and code generation that fit the context of what we're trying to do, and it can help us with debugging and troubleshooting, as well as documenting our code and providing information for the next people who will come along to know what was put there in the first place. As a craftsman, you could say, maybe even as a tools evangelist, I've spent a lot of time exploring how to share what's possible with the tools that we use in our everyday life. For me, that was Visual Studio and now it's VS Code. Figuring out how to get the most out of these takes time and effort. Visual Studio introduced the idea of suggested completions with IntelliSense a long time ago, and it's worked tirelessly to improve it. Over the years it's gotten better, but with the release of AI, it's moved forward by leap years. The deep suggestions as well as the analytics for understanding the code have made the life of the developer so much better. And it's not just these tools. ChatGPT, just by itself, is capable of generating code from scratch. From a description of what I want to create, it'll go out and create actual code using a Codex. The better the description, the better the result. For example, I saw a conference where the presenter stood up in front of the audience and asked GPT to write a asteroids game, and it created an HTML file and JavaScript that implemented the game with moving asteroids and even the ability to move your ship and shoot at them. It was pretty basic, but it gave a good starting point. That opened my eyes to what's really possible with AI. ChatGPT also helps to debug errors and suggest fixes as we can use it to convert from one type of a framework to another. For instance, I could take that JavaScript version of the asteroids game and have it converted to something completely different, like maybe Ruby or Python. ChatGPT is not the only player out there. GitHub Copilot is a resident now in our development tools like VS Code and Visual Studio, and is constantly evaluating the context of our environment and providing suggestions on what we could do with our code. If I had a comment that describes that I want to deploy a resource group, it might go out and suggest a resource group deployment using the AZ command. Things like this just make the job of writing code a whole lot easier because it not only offers the inline suggestions and enhanced IntelliSense, but it can also be used to explain code and give us context on what's going on. Let's take a look at how these work in real-life and see what we can do.

Contents