From the course: Rust for Data Engineering

Introduction to the AI coding paradigm shift - Rust Tutorial

From the course: Rust for Data Engineering

Start my 1-month free trial Buy for my team

Introduction to the AI coding paradigm shift

“

- [Instructor] Here we have a new paradigm shift, which is AI pair programming, and it combines very well with existing best practices, including DevOps. Here we have the GitHub ecosystem on the left here, and there are many amazing things you can do, like have GitHub Actions, your CI/CD server that automatically does matrix testing or linting or builds packages for you or even compile software in the case of Rust. And then you have Codespaces, which is a triggered environment specifically set up for, you know, let's say Rust or set up for Python or set up for an existing version of Ubuntu. So you've got this containerized environment here that's really a production-first mindset. Also, you use a top editor like Visual Studio Code, which is one of the most popular editors that has all of these hooks and, you know, plugins inside that allows you to be very productive. Now, once you've got all that set up though, one of the things that you'll be seeing in the future is that you'll be using not just an AI pair programming tool like GitHub Copilot but you also could be combining this with other tools. For example, you may first ask questions about a particular problem with ChatGPT. Take that, put it into copilot and then finish it off. You may not like in some scenarios exactly what you got or there is a limitation. For example, the amount of text that you could put into ChatGPT. And so you could go to another tool like Google Bard and actually double check what it's talking about. Maybe even stick with Google Bard for a particular segment of your coding, maybe a 30-minute window until you get stuck. And then you actually could toggle to yet another tool. For example, AWS Code Whisperer. So I think we're going to see people using lots of different resources, just like they do before AI pair programming where you'd go to Stack Overflow, Google, maybe to a learning platform to take a look at certain things. You're going to be combining all those things together to actually get the best hybrid experience when you're doing peer programming. Now, also pay attention here that the DevOps infrastructure doesn't go away. In fact, it's even more valuable because it's the best practices that are automated so that when you do have a result from an AI pair programming tool, you can validate that those actions are appropriate. You can use linting formatting tools to clean up the results. And finally, if you're using infrastructure as code, it will programmatically deploy that to a particular environment. So really, it's not just a replacement to use these AI pair programming tools, it's actually a synergy with the existing best practices that you're using. And in fact, that's the best way to use AI pair programming is to have this best practices of continuous delivery, continuous integration and then enhance that with AI pair programming.

Contents