Introduction
In case you're not familiar, GitHub Copilot is GitHub's "AI pair programmer" that provides intelligent code completion and a chatbot capable of answering any programming-related questions.
The concept behind GitHub Copilot is to enable developers to generate code more quickly, primarily by minimizing the amount of typing needed for boilerplate code and saving time when searching for answers to questions. Instead of using Google to find a solution, you can simply enter your question into the chatbot integrated with your IDE and receive a detailed, customized answer for your program in seconds.
In this review, I will provide my honest feedback on my experience using GitHub Copilot in various contexts, such as learning a new programming language, learning and practicing machine learning in Python, and developing a personal project. I will discuss some of the standout features that I use, identify areas where I believe improvements could be made, and share my overall evaluation of Copilot.
To preface, although this is not a sponsored post, I am able to use Copilot for free as a university student. My access to GitHub's student developer pack, which includes a free Copilot subscription, is granted due to my student status. As a result, my review may be somewhat biased, as the average person would need to pay for a subscription.
In addition to the point above, as a student, my experience with Copilot may be biased because I lack professional experience as a software engineer. Consequently, I cannot attest to how Copilot would perform in a "real-world" setting.
My goal in this review is to offer insights into how Copilot functions, as well as its features and drawbacks, before someone decides to purchase a subscription for its use.
The Pros: Unveiling the Strengths
GitHub Copilot is an incredibly powerful developer tool that, at times, feels as if you have an additional developer by your side, attentively observing your every move and offering assistance. This unique characteristic can be both a strength and a weakness, depending on the specific situation. In this section, we will focus on the positive aspects and strengths of GitHub Copilot, shedding light on its capabilities and potential benefits for users.
As a cutting-edge AI-powered tool, GitHub Copilot brings a wealth of knowledge and experience to the table, assisting developers in writing code more efficiently and effectively. By providing real-time suggestions and recommendations, it can help streamline the development process, reduce the time spent on repetitive tasks, and even offer insights into best practices and coding standards.
Moreover, GitHub Copilot's ability to learn from a vast repository of code and adapt to individual coding styles makes it a versatile and valuable asset for developers across various skill levels and programming languages. This adaptability allows it to cater to the unique needs and preferences of each user, fostering a more personalized and productive coding experience.
In summary, GitHub Copilot's strengths lie in its powerful AI-driven capabilities, real-time assistance, and adaptability to individual coding styles. These features combine to create a tool that can significantly enhance the development process and help users achieve their goals more efficiently. However, it is essential to consider that this level of assistance can also present challenges in certain situations, which we will explore in the following sections.
Efficiency and productivity
One of the most appealing features of GitHub Copilot is how little code I sometimes need to type. In very simple scenarios, it can suggest precisely, or almost precisely, what I intended to type. Moreover, the suggestions aren't just brief lines here and there; Copilot's recommendations can span several lines and complete entire functions. Consequently, you naturally spend less time fussing with typing boilerplate code and code that feels rudimentary to type.
For example, I recently embarked on a learning journey through the free lectures hosted on FreeCodeCamp's website, aiming to complete the certification for "Machine Learning with Python". Throughout the entire process, I had Copilot as my trusty companion in VSCode. To my amazement, I found myself typing very little code, even when the lectures delved into more complex, code-heavy topics.
Since Copilot is trained on a vast array of code sources available online, I surmise that it may have been exposed to the code featured in these lectures, which also appears on the TensorFlow website. Regardless of whether this was the case, the experience of not having to type out any code and instead being able to concentrate on grasping the underlying concepts of machine learning was truly refreshing.
This was particularly beneficial considering that machine learning code often involves a significant amount of boilerplate code and intricate syntax that even seasoned developers aren't expected to recall by heart. In such scenarios, Copilot truly shines as an invaluable tool, enabling learners and developers to focus on understanding the principles and techniques of machine learning, rather than getting bogged down by the minutiae of code syntax and structure.
Consequently, you feel incredibly productive and efficient as a developer who is in the process of learning something new, since you can complete entire programs while also focusing on understanding the concepts.
Accuracy and usefulness
As previously mentioned, Copilot not only provides code suggestions at every step of the development process, but it also demonstrates remarkable accuracy in specific contexts.
For example, when I used Copilot while learning machine learning in Python, I was amazed to find that Copilot's code suggestions were accurate down to the finest details, even including the exact URL of the datasets I wanted to use in my programs. This level of accuracy is truly impressive and sets Copilot apart from other tools.
In my experience, I'd estimate that for basic tasks, Copilot's suggestions are accurate about 99% of the time. This makes Copilot an incredibly valuable resource for developers, especially when compared to traditional methods like Google searches, which don't always yield accurate results quickly. With Copilot, not only is the answer available in real-time, but you can also trust that it's highly likely to be accurate.
This high level of accuracy and real-time assistance can significantly improve a developer's workflow, reducing the time spent searching for solutions and increasing overall productivity. By providing reliable and contextually accurate suggestions, Copilot has the potential to become an indispensable tool for developers of all skill levels.
Time savings
Finally, let's discuss time savings. As suggested by my first point, Copilot saves a considerable amount of time that would otherwise be spent typing boilerplate code or basic code.
However, there's another significant area where Copilot demonstrates its ability to save time: documentation. By providing developers with accurate and contextually relevant suggestions, Copilot can help streamline the process of writing comprehensive documentation for software projects. This not only saves time but also ensures that the documentation is consistent, well-structured, and easy to understand.
Developers often spend a considerable amount of time researching and writing documentation to ensure that their code is easily maintainable and understandable by others. With Copilot's assistance, developers can focus more on the core functionality of their projects, while the AI-powered tool takes care of generating relevant documentation.
For instance, I recently utilized Copilot to rapidly generate documentation for several of my personal projects. I was genuinely impressed by the quality and depth of the content it produced in the README files. The AI-powered tool was able to understand the core functionality of my projects and generate relevant explanations, making it easier for others to comprehend the code and its purpose.
However, it's important to note that Copilot's documentation capabilities, while impressive, are not entirely comprehensive. This means that certain aspects of the projects were left unexplained or not covered in sufficient detail. As a result, developers may still need to invest some time in manually creating or refining documentation to ensure that all critical components and nuances are adequately addressed.
However, for kickstarting documentation, Copilot is a decent tool. I don't know about you, but documentation is typically my least favorite part of the development process, so I'll gladly accept any time savings I can get in this area.
The Cons: Addressing the Limitations
This wouldn't be an honest review if I didn't discuss some of the cons and drawbacks that I personally encountered, and which I believe others may experience as well.
I will discuss 3 major cons, but keep in mind that there are plenty more that I won't be mentioning, which have been previously addressed by others in their commentary.
It's important to acknowledge that AI tools are still quite new, and as a result, they're in their early stages regarding how advanced they are at accomplishing their tasks and goals. I believe that as time goes on and these tools continue to be developed, they will only improve.
I also want to preface by saying that I don't buy into the fear-mongering that tools like Copilot will take the jobs of software engineers and developers. As these tools continue to improve, I believe they will simply be integrated into the workflow for developers, making each developer more capable and productive. Ultimately, the primary goal of these tools is to enhance developer productivity, not to create tools that will replace developers themselves.
Ambiguous/incorrect code suggestions
Now, I understand that this point may appear to contradict one of the advantages I mentioned earlier, where Copilot provides highly accurate suggestions. However, it's crucial to acknowledge that I was referring to relatively simple contexts. In more complex situations, particularly those involving proprietary software or specialized tools utilized by businesses, Copilot's suggestions might not be as impressive or accurate as one would hope.
It's important to remember that AI models like Copilot are trained on vast amounts of publicly available code, which means that their understanding of proprietary or industry-specific tools may be limited. As a result, developers working with such tools may find that Copilot's suggestions are less helpful or even incorrect in certain cases. This is not to say that Copilot is entirely ineffective in these situations, but rather that its effectiveness may be somewhat diminished.
That being said, I cannot definitively confirm whether this is true or not, as I am still a college student without direct experience working in the field. However, it's worth considering this potential limitation when evaluating the usefulness of AI-powered code assistants like Copilot. As these tools continue to evolve and improve, it's possible that their understanding of specialized tools and proprietary software will expand, making them even more valuable for developers across various industries and domains.
Cost
This is both a drawback I have yet to experience (since I'm currently able to use Copilot for free as a college student) and a drawback that is not subjective, but simply factual: Copilot costs money.
At the time of writing this article, the cost for Copilot is as follows: $10/month for individuals, $19/user/month for businesses, and $39/user/month for enterprises.
Now, aside from the fact that Copilot would be another monthly subscription that might catch you off guard, you have to ask yourself: "Is it worth the cost for what I do?".
I cannot answer this question, as everyone's situation is unique. However, I will say that if I were a student and had to pay $10/month for Copilot, I wouldn't. In my opinion, unless your company provides Copilot as a complimentary "perk," it's not worth using.
Although Copilot is impressive, there are plenty of free and user-friendly resources available, including both traditional materials and AI tools like ChatGPT.
If Copilot's capabilities significantly surpass the free resources available, then perhaps I would reconsider my stance on paying for it.
That being said, GitHub does offer a free 30-day trial for trying out Copilot. So, if you want to see what all the hype is about for yourself, you can at least do so without any cost.
Annoying while learning new language
Unlike the other drawbacks I mentioned, this one I can personally vouch for.
One of the use cases where I've tried Copilot is while learning the C programming language. I quickly realized how much it hindered my learning, so I disabled it in my workspace designated for learning C (which is very easy to do, by the way).
Recall at the beginning of this article when I mentioned that Copilot is like having another developer constantly looking over your shoulder and providing commentary every second? Well, this isn't ideal when trying to learn a new programming language.
When embarking on the journey to learn a new programming language, I found that Copilot transformed from a helpful assistant into an intrusive companion, negatively impacting my ability to concentrate on the learning process. The primary objective of learning a new programming language is to gain a comprehensive understanding of the syntax, enabling efficient navigation and problem-solving within the language. Additionally, it's crucial to develop the skills required to independently create and manage projects using the language. However, with Copilot's constant input and suggestions, it becomes increasingly challenging to achieve these goals.
As a developer trying to learn a new language, it's essential to engage in hands-on practice, trial and error, and problem-solving exercises to build a strong foundation. Copilot's continuous presence and guidance can inadvertently lead to over-reliance on its suggestions, preventing learners from developing their own understanding and intuition. By constantly providing commentary and solutions, Copilot may hinder the growth of critical thinking and problem-solving skills, which are vital for mastering any programming language.
In summary, while Copilot can be a valuable tool for experienced developers working on complex projects, it may not be the ideal companion for those attempting to learn a new programming language. Its constant input can obstruct the learning process, making it difficult for learners to develop the necessary skills and understanding to become proficient in their chosen language.
Conclusion
In conclusion, GitHub Copilot is a powerful AI-driven tool that offers a wealth of benefits for developers, particularly in terms of efficiency, accuracy, and time savings.
However, it's not without its limitations. The effectiveness of its code suggestions can be compromised in unique scenarios, the cost may be prohibitive for some, and its constant input can be more of a hindrance than a help when learning a new programming language.
As with any tool, its value will largely depend on the specific needs and circumstances of the user. Despite its drawbacks, Copilot's potential to enhance developer productivity is undeniable, and it will be interesting to see how it evolves in the future.
If you have any lingering questions or suggestions for an article you want me to cover in the future, feel free to leave a comment in the comment section below.
Lastly, make sure to follow my newsletter so you never miss out on when I post new content! When you subscribe to my newsletter, you'll be able to read my articles straight from your inbox as soon as they're released.
This article is part of a series called Bit by Bit, a series devoted to all things programming. Whether you're still a computer science undergrad or the CTO of Apple, there's something for you here.
New articles in this series are posted every Tuesday!
