How do Visual Studio Code and GitHub Copilot differ?
The Harvard Chan AI Glossary
In our classes, we often pause to explain the ideas behind AI and the language used to describe it. These posts grew out of those conversations. They’re for anyone who would like to understand AI better, with no technical background assumed.
Visual Studio Code, often called VS Code, is a code editor and development environment. GitHub Copilot adds AI assistance in supported working environments, including VS Code. They are distinct products: the editor is the workspace, while Copilot provides AI-supported capabilities within or alongside that workspace.
Code is the set of instructions a computer executes. The editor lets a person navigate files, read and change those instructions, and connect to tools for running or checking them. Extensions are additional packages that add capabilities to an application. AI assistance can be one such capability, but an editor also has many non-AI functions.
Copilot can support activities such as suggesting code, answering project questions and carrying out agent workflows. An agent workflow allows software to choose and execute some steps toward a goal. The selected model—the trained component generating responses—is another layer of the arrangement, distinct from both the editor and the assistant product.
Consider a learner improving a website’s search function. The editor displays the files. The AI assistant proposes or applies changes. Tests check whether specified searches return the expected results. The fact that these activities appear in one window does not mean one component is responsible for all of them.
Permissions determine which operations the assistant can perform and when a person must approve them. Access to project files does not automatically imply permission to publish a new version or alter unrelated records. The configured environment and organizational rules matter.
You might use an individual account or access provided by an organization. Those arrangements can differ in available features, model choices and controls. A glossary cannot infer every user’s capabilities from the product name alone.
The practical distinction is useful when troubleshooting or comparing tools. A problem may come from the model’s response, the assistant’s tool use, an editor setting or the application being built. Naming the editor, AI assistant and model separately makes the system easier to understand. The combined workspace can accelerate work, but the resulting behavior still needs review and testing.
Source: Microsoft — Build with AI in VS Code.
If you would like to find out more about AI and improve your leadership skills, consider the AI in Health Care Certificate of Specialization from the Advanced Learning Academy at Harvard T.H. Chan School of Public Health. You can take an individual course or work toward the full certificate.
AI is a fast-moving field. This explanation reflects our class discussions and related material at the time of writing. Terminology, capabilities and tools may change.
For suggested additions or corrections, please email me at tpanch@hsph.harvard.edu.
Dr. Trishan Panch
Program Director, AI in Health Care
Harvard T.H. Chan School of Public Health — Advanced Learning Academy