AI will make our jobs easier and not replace them
I firmly believe that AI will make our software developer jobs easier and not replace them. In this blog post, I will talk about the fear and uncertainty around AI and how it will change our jobs.
Table of contents:
- ChatGPT-4 is here
- The end of front-end development
- Ways to integrate AI into your developer workflow
ChatGPT-4 is here
ChatGPT-4 got released recently, and everyone is talking about it. OpenAI says it is more accurate, creative, and collaborative than the previous iteration ↗.
Watching this evolution of AI and how it might change everything in such a short time is mindblowing. But on the other hand, many people are worried about what will happen to their job when AI can do everything better and faster than humans.
New research ↗ by OpenAI, Open Research, and the University of Pennsylvania has revealed which jobs are most at risk of being automated by AI.
I understand the concerns around this topic. However, it seems like AI might already replace illustrators because so many startups get created which offer their users image generation services.
The end of front-end development
In addition to his, I stumbled over this article by Josh Comeau, the end of front-end development ↗. Although it's a very clickbaity title, it's a very interesting read, and I agree with the author and share his beliefs.
But, contrary to the title, the author explains that AI won't be the end of front-end development as we know it.
He shares several reasons why AI won't replace developer jobs. I've already talked about most of them with other developers, friends, and colleagues.
It's so funny that I wrote about this in 2021 when GitHub Copilot was released. And it's interesting to see that my beliefs haven't changed. You can check it out here: GitHub Copilot is not stealing our jobs.
First and foremost, it's astonishing to see that ChatGPT can build simple websites from a hand-drawn sketch ↗. But, as Josh also puts it, that's a job we've seen being done by other tools already for a long time. To build such simple websites, so many tools have been available for years, yet web developers continue to exist.
I'm currently working on a large project which includes two different APIs and several services in separate repositories. I'm primarily working on feature improvements there and fixing bugs stretching through several services.
I always think about how AI will be able to do this job for me. We're talking about hundreds of files and thousands of lines of code. So there's a huge difference between generating simple apps and websites and working on a large project with many services.
So I don't think AI can do my job soon. We're a very long way from that.
To put this more into perspective, a recent study found that ChatGPT4 produces misinformation more frequently and more persuasively than its predecessor ↗. so there is still a lot of room for improvement.
But I'm also not scared at all. On the contrary, I'm very excited about the future of AI and how it will change our jobs. For example, I think there will be more and better ways to integrate AI into our work which will lead to increased productivity and value.
Josh also states in his article that we might delegate tasks to AI, but we will still need to understand how it works and how to use it, and I agree.
Ways to integrate AI into your developer workflow
To summarize the thoughts from above, I think it's time to be excited about AI and not be scared and fear the loss of our jobs. To encourage this, I want to share a few ways how we can integrate AI into our developer workflow already.
GitHub Copilot is a great example of this. It's already two years old, and many developers have integrated it into their workflow.
If you haven't heard of it, it's an AI-powered coding assistant that can be integrated via an extension into your IDE, like VSCode. It can provide code suggestions based on your already written code or write the complete functionality of a function for you if you give it a descriptive name.
It was free at first but now costs around 20$ a month, but I think it's worth it.
It helps me to increase my productivity. It provides me code solutions matching the context I'm in and based on the already written code. The best thing is: it writes tedious and repetitive code for me. As you write there's always grey suggestions which you can enable by pressing tab.
It's also great with helping to reduce errors because it provides you with the correct syntax and knows the correct names of functions and variables.
I can't imagine working without GitHub Copilot anymore. I always think about a problem, figure out a way to solve it, and write out the base layout and skeleton. GitHub Copilot then fills in the details, which I then tweak to my needs and preferences. That's a perfect symbiosis between AI and me as a developer. I give the direction and figure out the architecture, and AI fills in the blanks.
Besides GitHub Copilot, ChatGPT can also be integrated in so many ways into your workflow already to increase your productivity and value created.
One example is to use it to write tests. You give it the code of a component, for example, and it will figure out every behavior and scenario this component is in and write useful tests for it.
Another use-case is to use it to write documentation. I haven't tried this yet, so I'm not sure how well this works, but from what we've seen already, it should be very helpful with this. You give it examples of your code and it will write the documentation for you.
AI-powered tools like ChatGPT can also be integrated into your developer workflow when you try to learn a new language. For example, imagine you're a front-end developer who works on a project where the backend is written in Go. You'd like to get a better understanding of a specific backend service because you want to work with the data that gets manipulated there. But you haven't used Go much before.
ChatGPT can help you to understand the code and the concepts of Go and answer your questions. So it helps to increase your productivity, and you won't need to bother your colleagues with questions all the time.
It's also great to get a better understanding of certain concepts of a new framework or new additions to a language that you haven't used before.
Finally, I often use ChatGPT to write specific functions that I didn't know how to write. For example, recently, it wrote me csv to json compile function in node with typescript. I could've written it myself with some research, but it was faster to just ask ChatGPT to do it for me.
GitHub Copilot X
I already finished this blog post and published it, and minutes later, I saw the release post of GitHub Copilot X ↗. Wow! When I wrote this blog post, I thought about how cool it would be to have ChatGPT in your IDE. And now, with GitHub Copilot X we get exactly that.
A ChatGPT-like experience in your editor with GitHub Copilot Chat, which does far more than the already existing GitHub Copilot, which can only suggest code.
It enables you to have context aware conversations. It knows all the ins and outs of the codebase you're working in and can help solve bugs, improve features or help you in general when you're stuck somewhere. That's crazy.
I can't even imagine how the next few years will look like and change the way we work.
But GitHub Copilot X is not only a ChatGPT addon. It also implements AI improvements for your Pull Requests, to get AI-generated descriptions for them.
With GitHub Copilot X, you will also be able to get AI-generated answers about documentation. When you have questions about the documentation, it will provide accurate responses out of the official documentation of the technology. This will be available for the React, Azure, and MDN Docs first.
And finally, it will be integrated into the command line interface, to compose commands and loops to use for you in your terminal. Have you ever been searching the web to write a specific command to use in your terminal? Now you can just ask GitHub Copilot X to do it for you.
But: GitHub Copilot X is not the new version of Copilot and won't be available soon. GitHub calls it a representation of GitHub's vision for the future. So it's not an available product we can use right away.
They are testing and trying out ways to provide it to us in the future. But it's still amazing to see what's possible and what's coming. I can't wait to try it out.
To sum it up, I think it's time to be excited about AI, not be scared, and engage with the fear and uncertainty around it. Instead, I think it's time to embrace it and integrate it into our developer workflow to increase our productivity and the value created.
Nobody can predict the future, but I think it's a healthy habit to welcome a tool like this that brings a lot of change and make the best out of it.
So far, these tools are making our job as developers easier, which I like a lot. Maybe I'm entirely wrong about this, and in a few years, we'll all turn into farmers, and AI will take over our jobs. But as long as it doesn't happen, I'm happy to embrace the change.
Sources and further reading
- GPT-4 is OpenAI’s most advanced system, producing safer and more useful responses ↗
- GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models ↗
- The End of Front-End Development by Josh Comeau ↗
- GitHub Copilot is not stealing our jobs
- ChatGPT can build simple websites from a hand-drawn sketch ↗
- Despite OpenAI’s Promises, the Company’s New AI Tool Produces Misinformation More Frequently, and More Persuasively, than its Predecessor ↗
- GitHub Copilot X: The AI-powered developer experience ↗
I hope you enjoyed this post and learned something new. If you have any questions, feel free to reach out to me on Twitter ↗ or via Email ↗.
If you want to support me, you can buy me a coffee. I would be very happy about it!☕️ Buy me a coffee ☕️
I wish you a wonderful day! Marco