Google Colab is rolling out two Gemini features that let notebook authors shape how the assistant behaves and how it teaches. Custom Instructions let creators encode preferences into a notebook, while Learn Mode shifts the assistant from delivering finished code to guiding learners through problems.
The additions target developers, educators and students who want a more controlled learning environment inside Colab, the company said in an announcement.
Related reading
- Anthropic secures multi-gigawatt Google and Broadcom chip deal as annual revenue run rate hits $30bn
- Microsoft introduces Critique in M365 Copilot
- Nvidia says the open versus proprietary AI debate is the wrong argument
Custom Instructions are saved at the notebook level and can be toggled from the Gemini chat box, so authors can embed preferred coding styles, class syllabi or library choices directly into a notebook. "Learn Mode turns Gemini into a personal coding tutor that provides step-by-step guidance instead of writing code for you," Google said; the company adds that Learn Mode uses Custom Instructions to break down complex topics and explain underlying concepts rather than supplying copy-paste solutions.
Users can enable Learn Mode from the Gemini chat window in Colab. Google also provides example notebooks preprogrammed to start a chat with the Gemini Colab agent and guide learners through Python exercises. Custom Instructions travel with a notebook when shared, applying to subsequent Gemini chats for collaborators and readers.
The recap
- Colab adds Custom Instructions and Learn Mode to Gemini.
- Custom Instructions are stored at the notebook level.
- Users can enable Learn Mode from the Gemini chat window.
