Google appears to be developing two intriguing features for its Deep Research tool:
Show Thinking
A "show thinking" button has been spotted in the Deep Research user interface. This feature is currently powered by Gemini 1.5 Pro, which doesn't support thinking capabilities. However, its presence suggests that Google may be preparing to integrate more advanced reasoning abilities into Deep Research.
Users are anticipating an upgrade that would leverage Gemini 2 models to power Deep Research. The addition of this button hints at the possibility of Gemini 2-powered Deep Research being in development. This could potentially allow users to see the AI's reasoning process, similar to how the 2.0 Flash Thinking Experimental model breaks down its thought process.
Create Audio Overview
Another new feature spotted in the Deep Research UI is the "Create Audio Overview" option. This suggests that Google may be planning to integrate audio capabilities into Deep Research, similar to the functionality already available in NotebookLM.
If implemented, this feature could allow users to generate podcast-like discussions based on their research outputs. The audio overviews in NotebookLM create lively "deep dive" discussions between two AI hosts, summarizing key topics from the source material.
It's important to note that both these features are currently hidden from users, indicating they are still in development and not yet ready for public release. The timeline for their potential launch remains unclear, as Google has not made any official announcements regarding these specific features for Deep Research.
