Do you think it would ever be possible to render html or JavaScript directly in ChatGPT? Or is there too much of a security risk to allow that.
I know there is an IPython function that ChaGPT has access to but that doesn’t seem to do anything
They will probably add it soon since some competitors already do it. Meanwhile I’ve built my own tool for this, that will render, have versioning and sharable links - codegpt-16db3.web.app
It’s very early Beta, but I’ve already built few basic web apps/one pagers with it.
Good work, I’ve seen these type of extensions before! I’ll check it out sometime.
Would be cool if it could use Ipython to render js like google colab. Tbh GPTs kinda feel like the next generation of colab notebooks.
I work with enterprise systems, and we’re on the “DOS” era of the GPTs interaction.
What I envision is a dynamic-generated UI to handle inquiries or commands from the users.
For example, when asked to fulfill a quantity, an interactive UI would come, and the user would scroll his finger to a slide.
Or, when selecting quantities, buttons with suggestions would come, and at the end, the data would be presented as a summary in a friendly-user screen.
All dynamically generated by the AI engine, instructed (not coded) by us.
A text-only approach isn’t the best way to handle such work. People are lazy, and typing things is boring. Even though you can speak with, having a grid design is a much better approach for medium to large amount of data.
Hi Klavs, thanks for your sharing!
They will probably add it soon since some competitors already do it
Do you mind share these competitiors? I’m very interested in this, very appriciated!
Probably the best and latest one is Claudes - it’s still in preview so you’d need to turn it on. Although, some js apis don’t work e.g. audio playback
see here How do I enable the Artifacts feature preview? | Anthropic Help Center