It is evident that Canvas (@canmore) has the potential to move beyond its current status as an experimental feature. With the right improvements, it could evolve into a serious AI-powered development tool rather than remaining a novelty. To achieve this, version control and seamless integration with GitHub are essential. Without version control, users cannot revert to previous versions or selectively merge changes, making edits effectively irreversible. The lack of GitHub integration further compounds these challenges, making it difficult to maintain AI-generated content within repositories and forcing users into cumbersome copy-pasting workflows.
For Canvas to become truly functional, it needs a system to track and manage changes effectively. Users should be able to view their edit history, revert to previous versions, and compare differences. An intuitive interface for handling these modifications is crucial. GitHub integration would enable direct repository pushes, in-place repository cloning for seamless edits, and native pull request support, allowing AI-generated content to integrate smoothly into professional workflows.
Some may argue that GitHub Copilot or existing ChatGPT integrations in VSCode already serve this purpose. However, while Copilot provides inline suggestions, it lacks the structured editing and revision control that Canvas could offer. VSCode integrations, though powerful, rely on third-party implementations and lack full native support. A first-party AI-IDE within ChatGPT, where the AI understands the full scope of a project—including its directory structure and version history—would provide a more cohesive and efficient development experience.
Canvas needs to move beyond being a test feature. Without version control and GitHub integration, it will remain impractical for real-world development.