Google Colab Rolls Out New Learn Mode

Google has just announced the introduction of two transformative features for Colab’s Gemini integration i.e. Custom Instructions and Learn Mode. These updates are designed to give users more control over their AI coding assistant.

Instead of repeating your requirements in every prompt, you can now define specific constraints. For example, if you are an educator, you can instruct Gemini to always reference a particular syllabus or use specific libraries when explaining concepts. If you are a professional working on a niche project, you can tell the AI to prioritize certain coding standards or documentation styles.

The most powerful aspect of this feature is that these instructions are stored at the notebook level. This means that when you share a notebook with colleagues or the public, your personalized AI assistant settings travel with it, ensuring that anyone who opens the file gets the same tailored experience.

While AI is great at generating code, it can sometimes hinder learning by simply providing the answer without context. Google’s new Learn Mode reverses this trend. When Learn Mode is active, Gemini shifts from being a “code generator” to a “coding tutor.”

Instead of providing a block of code for you to copy and paste, it breaks down complex topics into manageable, step-by-step instructions. It explains the logic behind the code, the underlying programming concepts, and why a specific approach is being used. This is particularly valuable for students or seasoned developers exploring a new framework who want to build their skills rather than just finish a task.

To help users get started, Google has released a series of example notebooks, including Python string and list exercises, where Gemini is preprogrammed in Learn Mode to guide users through common challenges.