Staying Relevant as a Software Developer in the AI Era

As a software developer with many years of experience, I always tried to keep learning new things, watching video courses, reading articles, and practicing my skills. But with AI, it was not easy. I felt like people were talking in another language. The theory and academic parts of AI felt too difficult to understand for me.

Changing Focus After Layoff

Recently, I had more free time because I was laid off from my job. I thought maybe this was a good time to go back to learning AI. I started with materials made for developers like the fast.ai course. It was a bit easier to follow, but I still felt far away from fully understanding what was being taught.

Also, I realized that to understand AI better, I needed to know Python very well. So, I paused my AI learning and started a long Python course, which was a bit boring, but I knew it was necessary.

Learning from Scott Hanselman at .NET Zuid

Yesterday, I traveled far from my home to attend the .NET Zuid event because Scott Hanselman was giving a talk. I met Scott in 2019 at the same event (you can see a picture at the bottom of this “Arabic” article, and I took another picture yesterday too, which I’ll share below). Scott’s talk was really interesting, and as expected, AI took a big bite of that talk.

In the past, people said “software is eating the world”, but now it seems like AI is eating the world. I sometimes wonder if we should be excited or worried about that.

Scott had some clear opinions about AI, and I want to share them here:

  • AI won’t take your job. Capitalism will. He explained that AI itself won’t steal jobs, but companies might use it as a reason to cut costs and replace people.
  • Use AI tools to help you work better. Scott said he doesn’t think he’s an amazing developer, mediocre in his own words, but he has great tools, including AI, that help him work faster and better.
  • Big AI models are not efficient. Some AI models use a lot of energy, and it’s better to use smaller models that do exactly what you need. He gave an example: if you’re working with C#, use an AI model just for C#. If you need Python help, use a Python model. Also, today’s CPUs can handle small AI models locally, so you don’t need to rely on cloud services.
  • AI is not creative. Even though AI can create things, Scott says it’s not truly creative. He’s not a fan of AI-generated images because it only feels like creativity but isn’t real.

A Helpful Answer That Changed My Perspective

After the session, I went up to Scott and asked him a question. I told him that with almost 20 years of software development experience, I’m starting to feel like I’m becoming slowly obsolete as AI grows more powerful. I tried to learn AI but found it very hard to keep up with the way it is usually taught.

Scott’s answer was simple but inspiring. He told me not to worry about becoming a data scientist. Instead, he said I should focus on learning how to find the right AI models for the job or task in hand or just for the sake of learning and exploring, download them, and run them on my own machine and play with them. He also suggested I explore tools like Hugging Face, AI Studio, and LLaMA models. Scott said that it’s more important to know how to use AI tools than to understand every part of how they work.

My New Path: Becoming a Super AI User

After speaking with Scott, I realized I don’t need to be a machine learning expert. Instead, I can focus on becoming a super AI user. This means I will:

  • Find smaller, specific AI models to use for my work.
  • Run AI models locally on my computer, using the new features in modern CPUs.
  • Learn how to use AI tools like Hugging Face and AI Studio to make my work easier.
  • Keep improving my Python skills, but only as much as I need for AI tools, not to become a data scientist.

Conclusion

For software developers who feel overwhelmed by AI, I have some advice: you don’t need to study the deep academic side of AI to stay relevant. Instead, focus on how AI can help you in your daily work. As Scott Hanselman said, the tools you use can make a big difference in your productivity.

I’m excited to continue learning about AI in a way that fits my strengths as a developer. My goal is not to replace my coding skills, but to make them stronger with the help of AI. Maybe this new approach will work for you too!


Mo Kinawy and Scott Hanselman #2