Cannot be directly executed. The generated code is mostly just a fragment without a - complete environment, such as dependencies and so on.
Poor debuggability: In some cases, AI-generated code may be difficult to debug and test because it can be hard for people to trace errors and logical mistakes in the code.
Infrastructure issues: Non-business-related parts, such as HTTP ports, database access, etc.
Limitations of existing solutions: Currently, the main solution for AI-generated code is using REPLs but they have limitations such as lack of code editing and saving functionality as well as limited support for programming languages and frameworks.
What should runtime do?
Environment isolation: providing independent running environments for each unit, allowing multiple units to run simultaneously in the same process without interfering with each other.
Dependency management: through automated dependency analysis and management, developers do not need to manually install and manage dependencies.
Debugging support: by providing an interactive console, developers can easily debug and test their code.
Infrastructure support: provides support for basic infrastructure such as HTTP ports and database access, making it easier for developers to write and run business code.
