ChatGPT's Deep Research Mode Gets a Fullscreen Document Viewer

[MacRumors]

OpenAI has updated its deep research feature in ChatGPT with a fullscreen document viewer that can be used to browse its AI-generated reports separately from the chat window.

The built-in tool includes a table of contents in a left-hand column that can be used to jump to specific parts of the document, while a right-hand column can still be expanded to show a list of citations.

The deep research mode has also picked up a couple of new options to help users focus research. From OpenAI's release notes:

We're introducing improvements to deep research in ChatGPT to help produce more accurate, credible reports with greater control. You can now focus research on specific websites and a larger collection of connected apps as trusted sources.

A redesigned sidebar entry point and fullscreen report view make it easier to start, review, and manage research in one place. Create and edit a research plan before it begins, and track progress with the ability to adjust direction mid-run.

Released last year, the deep research feature takes extra time to break a question into parts, then gathers and compares relevant information from the web, cross-checks claims, and synthesizes the findings into a better-supported, coherent answer.

Users can monitor ChatGPT's progress in real time and adjust the research scope, or add sources while a report is being generated. When the report is finished, it can be exported from the document viewer in multiple formats, including Markdown, Word, and PDF.

OpenAI says the new viewer tool is currently rolling out to Plus and Pro subscribers, with users on its ChatGPT Go tier and free users set to receive the update in the coming days.

Article Link: ChatGPT's Deep Research Mode Gets a Fullscreen Document Viewer

 

[iMac The Knife]

"We're introducing improvements to deep research in ChatGPT to help produce more accurate, credible reports with greater control."

Yeah, sure you are!

 

[thefredelement]

I've only tried to use this two or three times and it never finishes. It's gone on for like 30 minutes and just stopped, I typed 'continue' and got nothing.

 

[Mitthrawnuruodo]

Looks a bit like something Claude has been able to do for a while...And it _could_ be useful.

But, as always with LLMs, it comes down to accuracy...and that is rarely good enough. So, while there - most of the time - can be efficiency gains by using LLMs in many places, you still need to spend a good deal of time verifying...

...and sometimes, when the LLM is at its worse, you spend more time tweaking and verifying than "just doing it yourself" would've taken.

 

[apple2me]

Reminds me of NotebookLM from Google.

 

[iCyprus]

Reminds me of NotebookLM from Google.

I use NotebookLM a lot lately. So practical and to the point.

 

[svish]

Good to see this. Will be useful. Looking forward to trying it out once it comes to free tier/all users.