Contents:

Free course: “Quick start in Python»
Learn moreIgor Kamyshev is a renowned figure in his field, possessing significant experience and knowledge. His work is focused on achieving high results and implementing innovative solutions. Thanks to his professionalism and commitment to development, Igor has become an authority among colleagues and clients.
In his work, Igor actively uses modern technologies and methods, which allows him to remain competitive in the market. He strives for continuous learning and improvement, which makes him an expert in his field.
Igor Kamyshev also actively shares his experience with others by conducting master classes and training seminars. His approach to work inspires many, and the results he achieves confirm his high level of professionalism.
Cooperation with Igor Kamyshev opens new horizons for business and helps to achieve your goals. His reputation and achievements speak for themselves, making him a valuable partner and specialist in his field.

An expert in his field has deep knowledge and experience, which allows him to provide high-quality recommendations and solutions. His competencies cover a wide range of issues related to current trends and best practices. Constantly pursuing his own education and participating in professional events make him a valuable source of information for clients and colleagues. He actively shares his knowledge through publications and speaking engagements, raising awareness in his field. Drawing on his practical experience, he helps clients achieve their goals by offering customized strategies and approaches. At Aviasales, he works on the web platform team, improving front-end workflows. He maintains a blog and Telegram channel dedicated to software development. Here, he shares useful tips, new technologies, and current trends in web development, helping developers improve their skills and simplify their tasks. Links play a key role in web navigation and are an important element of SEO. They help users find the information they need and facilitate transitions between pages, sites, and resources. Well-structured and high-quality links improve a website's visibility in search engines and contribute to its popularity. It is important to pay attention to both internal and external links. Internal links connect your website's pages, making it easier for users to navigate and helping search engines index your content. External links pointing to your website increase its authority and trust in the eyes of search engines. Link optimization, including using relevant anchor text and correct URLs, also helps improve rankings. Effective links are the foundation of a successful SEO strategy and help attract more target audiences to your website.
Copilot is an intelligent assistant that integrates with development environments (IDEs). This tool can convert English speech into code, provide helpful hints, leave comments, and generate function names. The system demonstrates best results when working with programming languages such as Python, JavaScript, TypeScript, Ruby, and Go, but it also supports many other programming languages. Copilot significantly simplifies the development process, increasing the efficiency and productivity of programmers.
GitHub developed Copilot, a neural network trained on the open source code available on the platform. This system can also learn from local files, expanding its functionality and making it more adaptive to developers' needs. Copilot helps speed up the coding process by offering smart suggestions and autocompletion, significantly increasing programmer productivity.
Copilot's creators assure you that your closed-source code will not be uploaded to GitHub's servers. I share this opinion, because if GitHub wanted to access your code, there would be easier ways to do so. The security of your data is important to the platform, and they take measures to ensure the confidentiality of your code.
Copilot relieves me of routine work.
I am part of a team that develops tools for programmers. My work involves backend and frontend development, as well as Bash scripting. I was lucky enough to be a participant in the Copilot testing program, which gives me access to its free beta version. My friends are envious, as their applications for participation have not yet been approved. This service functions as a plugin for Visual Studio Code, making it a convenient tool for developers.
In the early days of using Copilot, I was dissatisfied. I was worried that the hints would be cumbersome and would require rereading to check the code quality, which would lead to unnecessary energy expenditure. In my work, I strive to save energy and minimize mental effort.
Copilot has significantly simplified my work, freeing me from mundane tasks. I actively used Storybook, a tool that allows you to develop UI components in isolation and instantly see them in the browser. Storybook previously had a plugin called Storybook Addon Knobs, which showed element properties, such as color or button size. However, it was disabled and replaced by a new plugin with a modified syntax. As a result, I had to manually rewrite 200 files, a process that couldn't be automated using migration tools.
After processing the first 15 files, Copilot significantly improved its performance, allowing me to rely mostly on auto-complete. I simply pressed Tab and occasionally made minor corrections. Although I don't have a deep understanding of neural networks and how they work, it's clear that Copilot quickly mastered the task and was able to complete most of the tasks independently. This highlights the effectiveness of AI-based tools in automating routine processes.
Copilot is primarily useful for writing JavaScript and TypeScript. However, it demonstrates insufficient performance with Bash, as the neural network hasn't received enough data to train in this area. If you develop in Bash, using Copilot isn't yet justified.
TypeScript is a JavaScript extension that adds static typing. This feature provides improved hints in editors and enables code validation at compile time. Adopting TypeScript significantly simplifies development standardization, which is especially important for large teams. By using TypeScript, developers can reduce errors, improve code readability, and increase overall team productivity.
I occasionally write Go, but I can't evaluate the quality of Copilot's suggestions as I'm just starting to learn it. I was disappointed with the autocompletion for Scala. The language is quite specific and has always seemed like a puzzle to me that's more interesting to solve on my own. While it's enjoyable, I don't plan to use Scala in production anymore. It's too academic for practical use.
I tried Haskell, and my experience was even worse than with Scala. I have a hard time remembering all the contravariant types, monoids, monads, and other concepts. These difficulties distract me from completing my regular tasks.
Copilot has become an effective tool for working with Docker Compose. For example, after the "docker compose" command, it immediately offers to forward a port, which is a reasonable approach. While it doesn't always predict numeric values accurately, its recommendations are generally helpful and greatly simplify the setup process. Using Copilot in Docker Compose increases productivity and helps avoid common mistakes.
It has many use cases.
I'll demonstrate how Copilot helps complete functions, generate names, and automate routine tasks. I'll use my pet project on GitHub as an example.
Copilot automatically completes functions in TypeScript. For example, if you open a file and start typing "formatDate," Copilot will complete the entire function. To save the code, simply press the Tab key.
The service offers an intelligent function signature. When adding text, Copilot will not be able to predict the full function, but will try to complete its signature. For example, if I need to calculate the sum of several values, it will offer its own solution. If I make changes, the system will immediately provide alternative suggestions.
I prefer concise hints because they're easier to work with. I often press Tab to check if my IDE highlights my code. If it doesn't, everything's fine.
Copilot generates new functions in TypeScript. When you type "function," it automatically suggests a name and code based on context. For example, it might suggest "handleChange," which is a logical choice. Copilot often suggests good suggestions, demonstrating its effectiveness in analyzing adjacent code and context. This significantly simplifies the development process and increases programmer productivity.
Copilot works even in an empty file. If I create a new YML file and specify "services", Copilot automatically suggests the correct code. At the beginning, an error occurred - a red indicator appeared at the bottom of the window, signaling that Copilot was not working, but it was activated soon. This assistant is capable of generating entire files, leaving me only the task of setting the necessary values.
Sometimes Copilot starts to get stuck on a certain fragment of code, suggesting the same lines. This happens until the user intervenes and makes changes to the code.
Copilot generates functions of varying lengths. For short functions, it generally produces good results, but when creating more complex and lengthy functions, its suggestions can be less successful. For example, when asked to shuffle an array, it generated a working function consisting of seven lines. This highlights both the capabilities and limitations of the tool when solving more complex programming problems.
I'm adding tests for my "format_number" function to check that it works in the US locale. I'll formulate the test as "should format in US locale," and the wizard will automatically generate the necessary code. The result is a logical structure, comparable to how I would develop it myself.
Copilot's hints adapt to the specific project I'm working on. The neural network seems to analyze the context and adapt to it. However, if Copilot offers overly long hints, it's often easier to ignore them and write the code myself. This can save time and improve focus, especially when the hint doesn't suit my needs.
Copilot has four important downsides
In September 2021, I'm writing this article about Copilot, which was still in beta at the time. It has several obvious drawbacks. It sometimes conflicts with the integrated development environment (IDE), and there are also questions about licensing and the quality of the generated code. Despite this, Copilot is an innovative tool that can significantly simplify the programming process, helping developers write code and increasing their productivity.
Conflicts with the IDE can negatively impact developer productivity. Copilot sometimes disrupts the built-in hints offered by the IDE. I primarily work with strongly typed languages, and when I type a period, the IDE displays the available object fields. The built-in autocomplete activates after 0.2 seconds, while Copilot's hint only appears after 2 seconds. This creates a delay and hinders rapid coding, which can be a challenge for developers striving for high efficiency. Optimizing interaction with the IDE and using autocomplete tools must be balanced to ensure a smooth workflow.
Foreign Code in the Project. Copilot can add fragments of third-party code to projects. There are concerns that the service may use elements of software with a malicious license, which would require licensing of the entire project. However, Microsoft has officially stated that it assumes all legal risks associated with the use of such code. This statement helps developers feel more confident using Copilot in their projects, but it is important to be aware of the potential implications of integrating third-party code.
A tweet was posted about a user entering an SSH key, and Copilot suggested a key completion taken from another project. However, the author subsequently deleted the thread and clarified that the neural network suggested a random key. This incident highlights the importance of being aware of the potential risks when using AI-powered tools, especially in the context of data security and sensitive information. Users should be mindful of how they handle SSH keys and other critical security elements to avoid potential threats.
If Copilot suggests code that is likely written by a human, I would treat it with caution. There have been no such cases to date.
Unclosed parentheses can be a problem when programming. Modern integrated development environments (IDEs) automatically close parentheses, simplifying the coding process. When you add an opening parenthesis, the IDE automatically inserts a closing one, minimizing the chance of errors. However, if you use tools like Copilot and reject its suggestions, this can result in parentheses being left unclosed, forcing the developer to close them manually. While this may seem like a minor detail, such small issues can be annoying and distracting from the core development process.
Typing errors. I want Copilot to correctly interpret the type system in my programming language. It often suggests incorrect types, which leads to compilation errors. This creates additional development difficulties and reduces productivity. Better typing support is needed so that Copilot can generate correct types and minimize the number of errors in the code.
I will always use Copilot
After a month and a half of using Copilot, I have become so accustomed to it that it has become my standard code completion tool in the IDE. I often press Tab for quick code completion. Since I mainly write simple programs, Copilot significantly simplifies routine tasks, allowing me to focus on the end result and avoid repetitive writing of similar code.
The system certainly makes mistakes sometimes, and in such cases, code adjustments are required, but it saves the developer a significant amount of time. When the service goes down, I immediately realize how important it is to my workflow. If Copilot weren't free, I'd be willing to pay for its services.
It would be helpful if Copilot automatically detected data types, validated its hints, and suggested solutions in case of type mismatches. This would significantly increase efficiency and improve collaboration with developers. We hope that such features will be added in future updates.

