In this article, I will share my most frequently used features of GitHub Copilot.
I am not a full-time developer anymore, but I still spend some time fixing bugs or reviewing code. AI, and specifically GitHub Copilot, is a tool that helps me a lot in doing my work efficiently.
My perspective on AI tools in software development
AI is not the future, is the present!
In fact, only the most distracted developers are unaware that AI is already part of our daily routines.
Obviously, AI will not solve everything, it won’t do all the work for us, and it certainly won’t take all the jobs. However, it will help us improve our daily tasks, making them faster and better!
I use AI as an accelerator tool to complete repetitive tasks, generate expected content, reproduce common patterns, and explore alternative approaches. I believe AI tools can assist in completing tasks, but they cannot fully replace the developer’s role. You still need to understand the context, ask the right questions, validate results, and fix any issues.
From my perspective, AI tools help with 80% of the tasks, but you still need expertise to handle the remaining 20%.
AI Buzz…
Unfortunately, the “AI narrative” has been heavily used by marketing departments in recent years to sell their services and products.
This has significantly increased AI’s perceived value. Today, if you don’t use AI tools in your daily tasks, you risk being left behind!
In my opinion, AI is not a “new thing”, it has been around for many years. However, the development of LLMs (Large Language Models) and the “prompt experience” offered by tools like ChatGPT has changed the game. These advancements have brought AI discussions into social media and made it accessible to the general public.
What is Github copilot
GitHub Copilot is a tool developed by Microsoft that helps developers with their daily tasks.
I use GitHub Copilot with VS Code (a free code editor), where it is available as a plugin for Visual Studio Code.
Pros
- Learning
I learn a lot with GitHub Copilot. Its code generation feature introduces me to new ways of coding, different approaches, and programming language features that I might not have known before. - Acceleration
You can complete your code faster with AI suggestions, generate sample data to test your algorithms, and even ask Copilot to generate code through a prompt. - Handling Boring Tasks
AI can assist with documentation, add comments to your code, generate commit messages, and more. - Quality
The quality of AI-generated code depends on your own coding skills. However, for 80% of developers, AI can generate better code, suggest better naming conventions, improve formatting, and add comments that many of us usually skip.
Cons
- Business errors
While AI can already understand some business rules, adjustments are still required. One of the main downsides is that you cannot fully trust everything AI generates. You still need to validate the expected results for your business requirements. - Deprecated code
One issue I have encountered when using AI for coding is outdated library versions. Sometimes, I had to teach the AI to use the most recent versions and adopt new best practices. - Low level of knowledge
This can be seen as both a pro and a con. AI allows developers to create complex software without being experts, but the quality of the generated code is often lower. Since AI is trained on historical code, and around 70% of developers are juniors, a significant portion of AI-generated code does not follow best practices.
AI is not an expert
I have noticed this in my daily use of AI. When I need to generate code that considers advanced topics like performance optimization, AI’s approach is often inadequate, and its suggestions are weak.
Most AI users have likely experienced this lack of expertise, when trying to explore deep or complex topics, it is common to receive incomplete, vague, or even incorrect answers.
The features I use the most
Answer technical questions using chat
Use Copilot Chat to ask technical questions within the context of your project.
Code suggestions
Copilot helps you complete code snippets efficiently.
Generate code
Ask Copilot to generate code — just make sure to describe your requirements clearly.
Documenation
In this example, I use Copilot to add JSDoc annotations, making it easier to generate comprehensive project documentation in the future.
Fill commit messages
Sometimes, it’s hard to describe a commit properly. Copilot can suggest commit messages, making this process easier.
Solve errors
One of the most useful features, ask Copilot to help identify and fix code errors.
Generate css based on html
To speed up development, I sometimes ask Copilot to generate CSS based on existing HTML.
Generate html based on css
Similar to the previous feature, but in reverse, Copilot can generate HTML based on existing CSS.
Review naming conventions
Ask Copilot to review naming conventions to ensure consistency with your project’s patterns.
Create package.json
Copilot can generate project files based on existing dependencies. For example, it can create a package.json file with all required dependencies.
