Artificial intelligence in programming

1. New partner in the world of programming

Artificial Intelligence (AI) is revolutionizing subsequent industries — programming is no exception. AI-powered tools are becoming an indispensable part of the modern developer's toolkit. They support us at every stage of work: from writing code, through its analysis and debugging, to creating documentation. Just a few years ago this seemed unattainable, and today it is difficult for me to imagine programming without using artificial intelligence.

Currently, even people unfamiliar with programming can create applications (so-called vibe coding) — by generating code in a chatbot and pasting it into an AI-integrated editor. As programmers, we shouldn't worry about this. The demand for our skills will not disappear — the nature of the work will simply change. Instead of manually writing code, we will focus on designing architecture, analyzing systems, and verifying and improving code created by AI agents.

In this article, I will show how to effectively leverage the potential of AI using popular tools like Claude Sonnet or GitHub Copilot as examples.

2. Claude Sonnet: programming genius

One of the most important players in the AI assistant market is Claude, and especially its latest model — Claude Sonnet 4. Created by Anthropic, it stands out with its exceptional ability to understand context and generate high-quality code.

2.1. How to get started with Claude.ai?

Getting started with Claude is very simple. Just go to claude.ai and create a free account. The interface resembles a classic chat where we can describe problems, ask for code snippets, or translate complex topics in natural language.

2.2. File transfer and the magic of Artifacts

Claude goes one step further than traditional chatbots. One of its key features is the ability to upload files. We can attach source code, documentation, or even a database. The model will analyze the content and respond within the context of the provided data.

However, what truly sets Claude apart is the function Artifacts. After generating code, a website, or a diagram, Claude creates a separate, interactive window — a so-called "artifact". You can edit the code, preview the page, and easily copy the ready solution. This is a groundbreaking tool that transforms the chat into an integrated development environment.

2.3. New level: working with AI in the terminal

And what if we want to move this interaction from the browser directly to the terminal and give AI access to the entire project? Console tools come to the rescue, allowing you to use Claude's power directly on local files.

Claude Code works on Linux and macOS systems, as well as on Windows with WSL enabled. After installation, we can work with a simple chat in the terminal that analyzes the entire project, creates new files, and edits existing ones — without the need to manually copy data to the browser.

2.4. GitHub Copilot: Your intelligent copilot in VS Code

GitHub Copilot is probably the most well-known AI tool for programmers, created in collaboration with OpenAI. It works as an extension for editors like VS Code or JetBrains and becomes our daily "co-pilot" in writing code.

2.5. Live chat and tips

Copilot offers two main functions:

1. Live Code Suggestions: when writing code, Copilot analyzes the context and suggests entire lines or blocks of code. Just press Tab to accept the suggestion. It can generate repetitive code, unit tests, or entire functions based on a description. Additionally, it allows for refactoring and translating selected code.


2. Chat in VS Code: the integrated chat allows for direct interaction within the editor. In the chat tab, we can select a model, add files to the context, and converse with AI while working on the project.

3. How to optimize your work

1. claude.ai: the ideal place to start a project, conceptual work, and debugging. It's worth describing your problem or functionality accurately and including a few files for context.


2. GitHub Copilot (chat in VS Code): best for working on individual tasks — asking specific questions and making changes to the code. It is recommended to focus on one problem at a time.


3. GitHub Copilot (inline chat): works well for editing single functions or a few lines of code. Avoid selecting entire files — this mode offers the smallest operational context.

4. Other notable AI tools

In recent years, hundreds of tools assisting programmers using AI have emerged. In this abundance, it's worth noting:

1. Cursor: a code editor that was designed from the outset with AI integration in mind. It does not require installing additional extensions.


2. Windsurf: another modern code editor with built-in AI features.


3. Amazon Q: an alternative to Copilot, working similarly, suggesting code in real-time. Furthermore, it also offers code scanning for security

5. Summary

The hammer did not replace builders — just as artificial intelligence will not replace programmers. The key is to learn how to use it wisely. Simply being able to write a few lines of code is not enough today. The market is changing — we must adapt to remain competitive.

Key takeaways:

1. Claude Sonnet: a powerful model focused on solving complex problems and working with code. The feature Artifacts makes it unique.


2. GitHub Copilot: the most recognizable tool supporting code writing and analysis, ideal for everyday use.


3. Automation and speed: AI handles repetitive tasks and writing boilerplate code very well.


4. Support and learning: tools help debug, translate difficult sections, and suggest better practices.


5. Working at a higher level of abstraction: AI allows focusing on functionalities rather than individual lines of code.


6. Verification is fundamental: AI still makes mistakes. The developer's role as a reviewer and responsible for the final code remains crucial.