GitHub is expanding Copilot, an intelligent assistant for programmers, with the introduction of Copilot Workspace. Now AI can act as a full partner at all stages of development.
Ahead of the annual GitHub Universe conference in San Francisco, GitHub announced Copilot Workspace, a development environment that uses what GitHub calls “Copilot-powered agents,” Techcrunch reports. It is a software development environment that uses artificial intelligence to help developers at every stage – from idea generation to testing and running code.
Enhanced Functionality
Copilot Workspace is based on GitHub’s open source programming assistant, Copilot. But while Copilot focuses only on generating code at the developer’s request, Workspace extends the functionality to help at a higher level. That is, Copilot, in conjunction with an assistant based on artificial intelligence, turns into a more universal tool.
Support Throughout the Development Process
As Jonathan Carter, director of GitHub Next, explains, research has shown that the biggest hurdle for developers is often getting started and deciding how to approach a problem. Copilot Workspace is designed to eliminate this pain point and collaborate with developers throughout the code change process. It acts as an intelligent assistant that meets with the developer at the very beginning of a task and helps reduce the “activation energy” required to get started, and then collaborates on changes to the code. The assistant can suggest ideas and approaches in natural language, help select files and methods for implementation. We’ll keep you updated as the development progresses.
Expanding Usage and Potential
Copilot is already actively used by developers – more than 1.8 million users and 50,000 companies. But Carter sees potential for even greater growth through expanded functionality in Workspace and related products, such as Copilot Chat for chatting with AI about code.
In his opinion, artificial intelligence can radically change the workflow, becoming essentially a partner, and Copilot Workspace is simply another step in this direction, complementing existing development tools. In the future, a full-fledged AI environment will be able to solve a whole class of problems that developers are not coping with as effectively today. Considering that Workspace is based on one of the most powerful neural networks today, GPT-4 Turbo, we can expect serious potential for the new development environment.
Addressing Challenges and Future Plans
However, GitHub Copilot currently generates a net loss for the company of about $20 per month per user. In this case, some clients even cost $80 in losses. This follows from an internal GitHub report reported by The Wall Street Journal. It would seem that such an economy should force the company to curtail the project. However, GitHub is not going to do this and announced a new product, Copilot Workspace.
Of course there are risks. Research has shown that tools like Copilot can generate unsafe and suboptimal code, notes NIXsolutions. Therefore, human control is stated as a key option in Workspace, that is, developers must independently guide the AI and check the generated solutions.
GitHub has not yet decided on a business model for Workspace and plans to collect feedback on the preview version. However, it is clear that the company aims to turn the technology into a highly profitable mainstream product, despite the current financial losses from Copilot. We’ll keep you updated on GitHub’s progress with Copilot Workspace.