This week, OpenAI, valued at 157 billion dollars and having received 6.6 billion dollars in investment, is continuing to add new features to ChatGPT, which gained 50 million new users in two months. OpenAI announced its new interface called Canvas in a post shared yesterday. Canvas presents a new way to interact with ChatGPT, opening a separate window alongside the normal chat interface. This window provides a workspace for writing and coding projects.
Users can create text or code directly within Canvas. After this step, they can highlight sections of their work to prompt the model to make edits. Canvas was made available yesterday in beta to ChatGPT Plus and Teams users, and will be rolled out to Enterprise and Edu users next week. According to OpenAI, the canvas may also be available to free users once it exits the beta phase.
With editable workspaces like Canvas, users can correct any incorrect outputs from the AI chatbot, examine prompts, and fix issues without having to create entirely new pieces of code. Additionally, Canvas offers features similar to Anthropic’s Artifacts launched in June and Cursor, which provides coding support.
Key Features of Canvas
According to OpenAI product manager Daniel Levine, ChatGPT will open Canvas windows when it recognizes that a separate workspace may be beneficial for generating longer outputs or for complex coding tasks. Users can simply type “use canvas” or “start canvas” to automatically open a project window.
In a demonstration for reporters, Levine showed how to compose an email using Canvas. When users request ChatGPT to create an email, it opens in the Canvas window, where users can adjust the length of the text via a slider. Additionally, users can highlight specific sentences, request ChatGPT to change the tone of the text, or add emojis. Users can also ask ChatGPT to rewrite the entire email in another language.
Levine also conducted a demonstration focused on coding. After asking ChatGPT to create a Python API web server, the desired server appeared in the Canvas window. When the “add comment” button in Canvas is pressed, ChatGPT includes inline documentation to explain the code in plain English.
Furthermore, if you select a code section created by ChatGPT in the opened screen, you can ask ChatGPT to explain that segment of code. You can also pose questions related to the selected code segment.
In addition, it’s worth noting that ChatGPT now has a new “review code” button. This button offers suggestions for specific edits, which users can approve, reject, or modify as needed.
Finally, note that during the iteration process, ChatGPT saves different versions of the canvas. Thus, if you prefer earlier outputs, you can revert to previous versions. If you do not want any of your uploads to be used by OpenAI to train its models, it’s sufficient to go to user settings and disable the model training option.
The product team has enriched the Canvas features of the GPT-4o model using synthetic data produced by OpenAI’s latest model, o1-preview. Among these features, the model is capable of recognizing when to activate the document and how to adjust it.
Source
Arwen Volkov, A graduate of the University of St. Gallen in Switzerland with a degree in International Finance, Arwen specializes in sustainable finance and green investments. She began her career at an investment bank in London, where she developed financing models for environmentally friendly projects. Known for her analytical and strategic thinking skills, Arwen is a sought-after financial consultant. In her spare time, she mentors fintech startups, contributing to their growth strategies. She is also a nature enthusiast and an amateur photographer.
