Although both writing and coding modes offer the option to request edits online, the bifurcated user interface for Canvas is designed with an additional set of shortcuts for those focusing on AI-assisted writing and another for coders. In the demo, Levine showed how the writer’s shortcut could be used to condense the number of words on a canvas or attempt to do a “final polish” on the draft. He also used one of the most cheerful shortcuts to add a bunch of random emojis. On the programmer’s side, ChatGPT can add logs, comments, and attempt to troubleshoot issues on a canvas.
ChatGPT saves different versions of the canvas as you iterate, so you can revert to previous versions if you end up preferring that output. Writers who may be concerned about OpenAI using what they upload to train their model should go into their user settings and make sure “model training” is disabled.
By allowing ChatGPT to make edits and suggestions, OpenAI is blurring the line between authoring and word curation. As someone who works with professional editors on a daily basis, I’m skeptical that the beta version of Canvas matches their incisive notes and careful guidance. But for people who don’t have easy access to human writing partners, I can see how it would be beneficial to get synthetic notes on a composition about structure and content.
It’s worth noting that three people listed as “supporting leadership” on the canvas project are no longer with the company. Former co-director and post-training co-founder John Schulman left in August and now works at Anthropic, a rival artificial intelligence company. Additionally, former CTO Mira Murati and VP of Research Barret Zoph resigned from their positions a week before this launch. At a press event at the OpenAI office after the departures, current chief product officer Kevin Weil reaffirmed the company’s commitment to continuing to release software.
“I think 2025 will be the year agent systems finally hit the mainstream,” he says. The idea of an AI “agent” that can not only work on software tasks alongside you, but is also agile enough to be sent out into the digital wilderness to do things on your behalf, is at the same time the recent past. and the projected future of generative AI.
Last year, WIRED covered ChatGPT plugins that people could use for tasks like booking flights with Expedia or making a reservation with OpenTable, possibly a step toward more “agent” AI tools. However, the plugins were later removed and more limited custom GPT chatbots were released in their place.
With this in mind, the Canvas beta appears to be another attempt to augment AI models with more decision-making capabilities, which may lead to surprises. During one of WIRED’s demos, Levine highlighted a portion of the canvas, requested an edit, and ChatGPT subsequently made an inline change near the bottom, outside of his highlight. “The really interesting thing is that a lot of times, if you highlight a section, an edit will be made to that part,” he says. “But ChatGPT has the option to decide where to edit.”
The closest alternative to the OpenAI canvas tool available right now is probably Google’s Gemini integration that allows you to use generative AI within Docs. Chatbots are definitely not dead, but AI companies are now recognizing the limitations of the format and looking for ways to diversify their software to discover new and catchy user interfaces. Google recently received praise in tech circles for its entertaining AI podcasts; Even CEO Sam Altman praised the tool.
With billions of dollars of investment still flowing through Silicon Valley into AI companies, consumers can expect to see more of these structural experiments that build on existing tools, such as AI podcast hosts and AI document editors. , which will be released at a regular pace over the next year. . The chatbot race is far from over, and future iterations of the technology are likely to move far away from that monotonous chatbox and take a more multifaceted approach.