With focus on traditional Artificial Intelligence (AI) and Generative AI (GenAI) over the last 2 years, there’s been a parallel development of another technology known as “copilots“.
Copilots can be compared to the digital assistants that have been around for quite some time now but what sets them apart is their advanced capabilities.

A major factor driving this trend is the rapid growth and adoption of Large Language Models (LLMs). Although copilot technologies are currently in their early stages, they are poised to take center stage in the coming years.

In this blog, I’ll guide you through an understanding of co-pilot technologies. You’ll also explore the various co-pilots currently utilized in software development.

So, let’s get started!

What are Copilots?

While you can use LLMs like OpenAI’s ChatGPT, co-pilots are especially powerful when integrated directly into your coding environment, offering a range of advanced capabilities.

Harness GenAI’s capabilities to develop top-tier software products.

Download Guide

Now, let’s dive deeper into the various co-pilots used in software development, with a special focus on AI pair programming.

GitHub Copilot

GitHub Copilot is a GenAI-powered code completion tool developed by GitHub, OpenAI, and Microsoft. It helps developers write code faster and more efficiently by suggesting entire lines or blocks of code as they type. It is available at the individual, business, and enterprise levels and offers a free trial for individual users.

For any selected code block or line of code, GitHub Copilot can help you with:

1. Analysis and Explanation

2. Debugging and Code Fixation

3. Document generation

4. Test generation

Below, GitHub Copilot can be seen in action analyzing and explaining a block of code: 

Here are the steps for you to follow for setting up GitHub Copilot in your IDE:

1. Sign up or register for the GitHub Copilot service to get a free trial.

2. Install the GitHub Copilot extension in your IDE of choice, such as Visual Studio Code, Visual Studio, JetBrains IDEs, Neovim, etc.

3. For Visual Studio Code:

• Install the GitHub Copilot extension.
• Sign in to your GitHub account to activate the extension.
• Optionally, set up the OpenAI Copilot integration by providing your own OpenAI API keys.

4. To use GitHub Copilot in your coding workflow:

• Highlight any piece of code in your IDE.
• Right-click on the highlighted code and select the “Copilot” option.

This will indicate that the GitHub Copilot extension is successfully activated and ready for use.

Recent Landmarks:

When discussing copilots, we’re excited to announce that Ascendion, our parent company, has recently launched Digital Ascender called “AVA”, a tool with powerful capabilities.

This tool is currently in a testing phase and will soon be available for installation and complete integration into our organization’s local programming environment. It includes features such as code completion, documentation, bug finding, unit test generation, code optimization, and reverse engineering.

Here is an example of AVA generating documentation for a piece of code:

Here, AVA demonstrates its ability to optimize provided code:

Returning to the concept of GitHub Copilot, GitHub Chat offers convenience with commands.

For example, to set up a new project, you can use the /new command followed by a short description in natural language:

• /new react app with typescript
• /new next js app

For code blocks, you can use the /fix command, and for generating tests, use the /tests command.

In the image below, you can see Copilot Chat generating the existing project structure and suggesting relevant follow-up questions:

When asked to fix a code block, it debugs and suggests an optimal solution for you to accept or discard. Once accepted, it updates the current code with the corrected version.

The image below demonstrates it:

Note: You can also directly add files and feed them to Copilot Chat. These are just a few of GitHub Copilot’s capabilities.

Currently in June 2024, GitHub Copilot is available for $10 per month at the individual level, with different pricing for business and enterprise users.

Did you know?

Microsoft claims, based on their data, that GitHub Copilot enables developers to code 55% faster than they would otherwise.

So, soon, GitHub plans to adopt OpenAI’s new GPT-4 model, introduce chat and voice features for Copilot, and integrate Copilot into pull requests, the command line, and documentation to answer project-related questions. This expanded functionality is part of Copilot X, recently announced by GitHub.

CodeGPT

CodeGPT is a popular extension that leverages Large Language Models (LLMs) to enhance programming tasks with AI. It allows dynamic interaction with your work environment and can create AI Agents to enrich LLMs with context and connect you with critical areas of your business.

Want to use CodeGPT in VSCode? Here are the steps you can follow to do it:

1. Install the CodeGPT extension from the VSCode marketplace.

2. Sign in or register using the CodeGPT portal within VSCode to authenticate your account.

3. Configure CodeGPT:

• Select your LLM model provider in the settings.
• Enter your API key from the selected LLM model provider.

4. Adjust CodeGPT Chat Settings:

• Customize Provider Attributes.
• Set Token Limits.
• Refine Temperature Control.
• Manage Window Memory.

Besides VSCode, you can create a free CodeGPT Plus account and sign in/register using GitHub or other options.

This gives you access to their platform where you can search the marketplace and load agents specific to your needs, such as GPT-4, Coder Assistant Agents, and UX Designer Agents. It also supports various libraries for data science tasks and popular programming languages and frameworks.

Here are the several tools that it offers:

• Chat: Engage in AI conversations using your chosen provider’s models or agents on CodeGPT Plus.
• AI Agents Marketplace: Browse agents in the Marketplace and interact with them, ensuring you choose CodeGPT Plus as your provider.
• React Sandbox: Explore React components by interacting with them or uploading images for experimentation.

The extension integrated with VSCode provides code completion suggestions based on the developer’s input, covering variables, functions, and methods. It also offers features like Copilot to explain, refactor, document, debug code, and run unit tests.

One additional capability that CodeGPT provides is the Vision tool. This is a powerful tool that allows you to analyze images and extract meaningful information.

To leverage it, follow these simple steps:

1. Establish a connection with CodeGPT Plus or Ollama to access the Vision tool.

2. Upload an image in .jpg, .png, or .webp format to use as a prompt for the model.

3. Identify and locate objects within the image with the help of Vision. Get details such as the type of object, its position, size, and color.

Please find below the extension and its look and feel integrated with Visual Studio Code:

In addition to these, I’d like to introduce you to two other copilots that can simplify your tasks. Keep reading!

Testim

Testim is a platform that leverages AI to automate functional testing. It specializes in creating, executing, and maintaining automated tests, providing a simpler approach for QA teams dealing with complex scenarios.

Here are its key features:

• AI-Based Test Automation: AI-driven algorithms create robust and maintainable tests.
• Self-Healing Tests: Automatically updates tests to reflect changes in the application UI, reducing maintenance efforts.
• CI/CD Integration: Seamless integration with popular CI/CD tools like Jenkins, CircleCI, and GitLab.
• Multi-Browser and Device Testing: Enables testing on a wide range of devices and browsers.
• Smart Locators: Uses machine learning to improve the accuracy and reliability of element locators.

Here are the two primary use cases for Testim:

• Ideal for Agile and DevOps environments where continuous testing is critical.
• Suitable for teams looking to reduce the maintenance burden of automated tests.

Applitools

Applitools specializes in visual UI testing and monitoring. Its AI-powered Visual AI technology helps ensure that applications look and function correctly across different devices and browsers.

Here are its key features:

• Visual AI: Advanced visual comparison algorithms to detect UI discrepancies.
• Multi-Browser and Device Testing: Enables testing on a wide range of devices and browsers.
• Seamless Integration: Integrates with major CI/CD and test automation frameworks.
• Baseline Management: Manages baselines and provides tools to review and approve visual changes.
• Root Cause Analysis: Helps pinpoint the cause of visual and functional issues quickly.

Here are the primary use cases of Applitools:

• Ideal for teams focused on ensuring consistent visual quality and user experience across platforms.
• Organizations needing to validate UI changes rapidly in a CI/CD pipeline.
• GitHub copilot can also be used for testing with tools like jest to run unit tests

After reading all the above information about various co-pilots, you might be wondering “why not use LLMs directly,” right?

Worry not as I’ve drafted specific reasons for the same for you to understand the difference.

Difference between GitHub Copilot and GPT-4

Although multimodal LLMs like GPT-4 are valuable for software development, they serve a different primary purpose and focus compared to copilots, despite having some overlapping capabilities.

Here are some differences between GPT-4 and GitHub Copilot:

By now, I hope you have a clear understanding about copilots and their specific purposes.

While some copilots may be limited in their current versions, they are expected to offer a comprehensive set of advanced tools in the future. So, you can rely on them to seamlessly integrate into the workspace, accelerating the pace of development, significantly boosting productivity, and saving time and resources.

To know more about cutting-edge modern tools and technologies, reach out to us at Nitor Infotech.