Google has released new content for its Gemini assistant called Canvas — a split-screen feature that lets you chat to Gemini on the left and see your changes appear in real-time on the right.
The idea is to make editing and iteration a smoother experience — instead of scrolling up and down the chat to copy sections of output you’re not happy with, you can just highlight the text in question on the right and tell Gemini what to change. The assistant will then edit the specified section and update the document, rather than generating a whole new version or spitting out additional paragraphs you need to splice together yourself.
Asking an LLM like Gemini to make revisions to its responses can be a bit of a chore, so this will hopefully make the process less painful.
Canvas also works with programming projects, allowing you to view code on the right and chat with Gemini to explain, revise, and debug it on the left. It can also display your HTML or React code as a visual representation of your software, allowing you to preview what your email subscription form might look like, for example. When you request changes, the preview will update, allowing you to try out different ideas quickly and efficiently.
To try out these new features, you’ll need to be a Gemini or Gemini Advanced subscriber and click the Canvas button in the prompt bar. Google is marketing these updates as features for “collaboration” but just to be clear — it doesn’t mean collaboration with other people. The features are designed for you to collaborate with Gemini and, according to Google, “if you want to collaborate with others on the content you just made, you can export it to Google Docs with a click.”
The update also includes Audio Overview, a feature from NotebookLM that essentially transforms documents into podcasts. It’s similar to any summary and analysis generation tool in its purpose, but it presents the information in an audio format with two AI hosts holding a podcast-style discussion.
The feature has been popular with NotebookLM users who want to multitask while consuming information. To use it, upload your documents to Gemini and click the suggestion chip that appears.
