AI Leaks & Rumors 👀 - TestingCatalog (Page 7)

11 Sep 2025 · 2 min read

xAI readies background conversations so Grok can work while you chat

What's new and what's next? Now you can select Grok mode on Projects; Soon it might be possible to select Grok model explicitly as well as run thinking tasks in the background within the same conversation.

10 Sep 2025 · 1 min read

Microsoft tests voice mode on the Copilot home screen and private chats

What we know so far: It will be possible to chat with Copilot via Voice straight from the home screen, without entering a voice mode. Private chats are also being developed.

9 Sep 2025 · 2 min read

Perplexity Comet to get local MCPs unlocking file access and app control

What we know so far: Users will be able to add custom Connectors to Comet, letting it access local files and control other desktop apps.

1 Sep 2025 · 2 min read

Microsoft prepares Copilot Appearance to a broader launch

What we know so far: Copilot Appearance is being prepared for a broader launch and is likely to get a new color. More developments are surfacing around the Memory feature and overall page layout.

1 Sep 2025 · 2 min read

Google tests new NotebookLM voices and audio formats Brief, Critique, Debate

What we know so far: Three new Audio Overview formats are coming to NotebookLM - Brief, Critique and Debate. New formats are also featuring new male and female voices.

31 Aug 2025 · 1 min read

Google develops AI-suggested Report options for NotebookLM

What we know so far: Studio UI in NotebookLM is about to get a set of upgrades to make artefact customisation more extensive. The "Reports" option will get AI-suggested recommendations based on the uploaded content.

30 Aug 2025 · 1 min read

Microsoft tests new categories for upcoming Copilot library

What we know so far: Copilot library is about to get new categories for podcasts, documents and quizzes!

30 Aug 2025 · 2 min read

OpenAI tests new tools for effort control and conversation branching

What we know so far: OpenAI continues its experimentation with the effort selector UI, as well as a new functionality for branching chat conversations.

27 Aug 2025 · 1 min read

Anthropic developing Claude Code web version to rival Codex

What we know so far: Claude code on the web will work quite similarly to Codex, where users will be able to connect their Github repositories and spawn Cloadue Code agents in virtual environments.

26 Aug 2025 · 1 min read

Google tests ingredient sharing and print ordering for Whisk

What we know so far: Whisk to get ingredients sharing, so users will be able to remix them. Besides that, a Print ordering button may appear in the Gallery soon.

24 Aug 2025 · 2 min read

Google tests new Gemini modes including Agent, Go and Immersive View

What we know so far: Agent Mode will power autonomous exploration, planning and task execution. Gemini Go will help with ideating and prototyping. An Immersive View will bring Visual Answers.

24 Aug 2025 · 1 min read

Microsoft testing new Study and Learn mode in Copilot

What we know so far: Study and Learn mode is coming to Microsoft Copilot ahead of the school season; It will likely mimic a similar feature on ChatGPT.

24 Aug 2025 · 2 min read

xAI tests new object highlighting feature in Grok voice mode

What we know so far: A new toggle called "Objects Highlighting" has been added to the latest build of Grok on iOS; Grok will soon be able to highlight objects in the camera view. Earlier, Elon Musk mentioned that Grok 4.20 is expected to arrive with advanced video understanding;

23 Aug 2025 · 1 min read

Google tests Gemini Canvas export to Firebase Studio for faster prototyping

What we know so far: Model selector will be moved to the prompt bar; A new option to export into Firebase Studio will surface on Gemini Canvas.

21 Aug 2025 · 2 min read

NotebookLM to pull sources from web and Google Drive via Deep Research

What we know so far: The Deep Research feature from Gemini is coming to NotebookLM. Likely with a possibility to pull sources from Google Drive as well.

21 Aug 2025 · 2 min read

Flow to get text to image generations powered by nano-banana

As anticipation grows around the nano-banana model, Google is steadily laying the groundwork for its integration across several creative tools. The upcoming deployment of this model isn’t limited to

20 Aug 2025 · 1 min read

Mistral AI prepares Memory feature for the Alpha rollout

What we know so far: A new announcement banner has been added to the Mistral AI on the web, announcing an upcoming Memory feature.

19 Aug 2025 · 1 min read

Perplexity prepares GPT-5 Thinking for Max subscribers

What we know so far: GPT-5 Thinking will initially be available to Max users and appear in the list of reasoning models.