ChatGPT Canvas is amazing!
Being able to delete documents is a must, it’s a must-have feature, folks.
Better organization is needed as well. Being able to navigate between documents without having to close a document, then open the list of documents, then click on another document. This is annoying and inefficient.
Also being able to view two or more documents side by side would be terrific.
Yes! And also, in a canvas chat:
And a much more advanced one, for later maybe:
I know ChatGPT can access your Google Drive files to load them. But that would be even better if it could write to it. This way, a document in canvas could be linked to a doc in G Drive that would automatically synchronize! That way, no more copy-pasting or download-uploading needed.
When GPT understands the file you are speaking of, but updates or overwrites the file in canvas instead of switching and making the change there. This creates project corruption. yes to deleting, yes to renaming, yes to updating canvas source with what you have in your IDE that is working.
Exactly! I made SO MUCH PROGRESS on my book today, it was literally the most helpful tool I’ve ever seen in terms of brainstorming and organizing my thoughts. But once ChatGPT canvas starting creating multiple documents, the guy had zero idea how to make the change I was asking for. NO, ChatGPT, I need you to make changes to Act 3!! “I tried making changes to Act 3, but I couldn’t” - also him, “Editing Act 1”.
I would also like to see these features.
We don’t necessarily need larger canvasses (also an issue when you run out of room) if we can organise the document list into chapters, scenes, etc. Although it seems to suit multiple chapters, rather than less chapters with more scenes (where each chapter is a ‘theme’ or ‘Part’). The ability to rename a Canvas, and delete old or unwanted Canvas’s would be useful. Currently, in my world building chat, I have multiple pages dedicated to one thing that could be made into a single page. I realise the naming issue is one I’ve created myself but still, things change.
It’s hard to imagine that the feature was rolled out without rudimentary organization features like delete and rename.
This is especially bothersome for writers. We frequently rename things during creativity, and these unused canvases can accumulate rapidly.
I have tried a workaround of creating a canvas called “Characters” and adding them all to that, but that’s a lot of text to scroll through, and ChatGPT runs into problems updating longer documents.
I like Canvas, but I’m disappointed that rudimentary features like this were not included in the release. These limitations make it almost impossible to use Canvas to organize a novel meaningfully.
The lack of a delete option is an enormous issue. I have been trying to work with Canvas and it has corrupted my work repeatedly by making duplicate canvases, with no way to get rid of the old ones or even tell them apart based on the name.
ChatGPT claims it can delete the extra canvas if I confirm… I confirm… then it informs me it can’t delete the canvas, but I can delete it manually. It gives me imaginary delete instructions that don’t work, either.
Is there any way at all?
I can relate to these replies. And to add that the support team denies the issues by blaming internet connections, cache, history etc. Its got so bad that I’ve gone from working on a whole book / document to just working on one paragraph at a time. Chatgpt Canvas is doing to my project - organisation ops included, what everyone else in these replies is experiencing. I still find it has poor memory in the Chatgpt Canvas too.
My strategy has been to attempt to avoid the corruption and overwriting, not implementing updates improvements etc, because I think its due to too many Canvas, (that Chatgpt AI hallucinates that I can delete them) that I don’t need anymore, being active in the Chat, that I’m now attempting to work in separate Chats, one Chatgpt Canvas at a time to keep my work safe. Its much slower, which then makes me think…meh, still progress is better…er, just not as efficient. I like many want these issues resolved.
I think I’ve found my issue. My laptop doesn’t run nor have a GPU, so using Chatgpt Canvas will have a lot of limitations for me till I remedy that situation by buying a PC. In saying that though there are plenty of work arounds, disconnecting devices to the internet, using Firefox browser.
I have this exact same issue. I’m glad it’s not just me! My documents aren’t even that big but it’s still throwing a fit…
I’ve stolen your suggestion. It’s taken me a full day to get everything moved and organised. I don’t like that it struggles to reference, and it can’t seem to be able to put thing in to chats for you. So I’m copy pasting, which is, like you say, slow. But at least it means I’m not losing anything and I can track my changes. I have a living document in each canvas for now. It’s better but I was making such fantastic progress on my project. Until it died on me.
There’s so much potential here though, it shouldn’t even be that hard to fix and implement real, worthwhile changes. At that point, why would you even use anything else!?
@gmw_158 Great stuff. Glad to hear your making progress. That’s all I ever want too. I plug directly into my modem as well whenever I use Chatgpt Canvas. On a PC the GPU makes a massive difference in timings for when prompts are sent and received by the model. With the GPU there’s no interruptions and therefore no faulty replies, responses and output from Chatgpt Canvas. But my laptop is all I got for now. Happy Outputs.
That’s an interesting approach. Can you elaborate a bit more? I work in tech and I’m interested to know your thought process. My router heads in to a switch that my pc plugs directly into. If I’m copy pasting I use my pc, it’s just easier. I use the app on my Samsung galaxy for general brainstorming. I find it a lot faster to get my ideas out using my phone. The app isn’t great though. I’m seeing way more issues when using is vs my pc.
Excellent for your PC you must get really good output from the model. I’ve only got a laptop intel core i7 that I run Chatgpt 4.0 and Chatgpt Canvas on. I do have the app on my Samsung as well. But the lag is too long to make it a reliable source to brainstorm with and it struggles to connect to the Chatgpt 4.0 account on the cloud, again that’s all wifi and internet infrastructure. The PC that I had the chance to experiment with, had a GPU with 16 GB of VRAM I think it was GeForce. The difference in use with prompts and commands was like night and day, the output was spot on without delays and without hallucinations - referring here to comparisons between my laptop and that high performance PC. I’m still thinking of how to build the environment to do two things, learn how to use AI and have an infrastructure that keeps that learning and usage sustainable. It’s out of my budget range for now. I hope that answers in part your question?
Bump - has anyone solved this yet? Or is this not a feature available yet?
Not solved yet. Struggling to get rid of empty document from my workspace 