Brittbot

Upgrading from Data Center Tech to Software Engineer

A black woman sitting at a desk in a cubicle in an office typing on a keyboard. A computer monitor displays lines of code.

Becoming A Junior Developer Who Can Code

I finally read the viral blog post that came out almost a month ago about how the newest junior developer can’t code. The author, using his own experience at his current place of employment, wrote about his displeasure seeing how these developers rely too much on coding tools like Cursor and/or Copilot. Others use ChatGPT or Claude to get their code snippets and perform their job. As a budding developer myself I rather not develop a crutch on AI tools. Thus, I avoid using those tools as much as possible. Here’s how I’m becoming a junior developer who can code.

Briefly Explaining The Author’s Position

The way I understood the author’s position in his post is like this:

  1. He doesn’t like how the junior developers he works with rather type in a prompt to ask for the code fitting certain requirements instead of trying to code it themselves.
  2. When these developers run into an issue with the AI-generated code, the author doesn’t like how they used the AI tools to seek a solution instead of trying to solve the problem themselves.
  3. Thus, these developers avoid the struggle of writing and/or debugging code so they don’t learn. Hence, they don’t know how to code.

How I’m Becoming A Junior Developer Who Can Code

First, when it comes to my Integrated Developer Environment (IDE) I don’t use any AI assistants like Cursor or Copilot. Of course I get notifications from my IDE trying to persuade me to use them (PyCharm and VSCode), but I ignore them. What I do instead is design my application using pen and paper. I write out the logic flow and a little pseudocode. Designing this way really helps me understand my objective, find any potential problems, and create solutions.

Second, I write the code, test it out, and debug it if it fails. If I mess up my logic, then writing it out again as laid out in the IDE to compare it to my initial design really helps me find the logic flaw. Sometimes I’ll talk out loud as that helps me too.

Third, if I get stuck during the debugging process I use Google Search for a hint. Usually that helps me get going as I can figure out the rest of the solution to my problem. Then it’s back to writing code, testing it out, and debugging if needed.

Finally, if the hint doesn’t help me I’ll use Gemini and ask it to explain to me why my code doesn’t work. When it provides a working solution I read through the explanation to learn where I went wrong. If the explanation confuses me I ask Gemini to explain it a different (or simpler) fashion. It usually does.

Conclusion

Becoming a junior developer who can code is fun, but also frustrating. The struggle to write functioning code, or debug non-working code, is tough. Sometimes it just suck. However, the feeling of finally fixing it makes the struggle worth it. And in the future I’ll know how to resolve that issue faster because I already dealt with it. By allowing an AI assistant or tool to do my work for me only cheats myself of becoming a competent code.