OpenAI is rolling out Group chats in selected countries across all plans!

“Rolling out now to Free, Go, Plus, and Pro users in Japan, New Zealand, South Korea, and Taiwan on both mobile and web.”

World Labs launches Marble 3D world model for creators

World Labs has launched Marble, a generative multimodal 3D world model now publicly accessible. It accepts diverse inputs, supports detailed editing, and offers export options compatible with standard 3D workflows across multiple industries.

🗞 #ai

ICYMI: Flowith AI releases FlowithOS app in public beta

Flowith AI has launched FlowithOS in public beta, a cross-platform desktop app designed for agent-led workflows. It combines browsing with task execution, structured action logs, and file-based note storage for productivity-focused users.

🗞 #flowith

LMArena launches Code Arena for full-cycle AI model evaluation

Arena has launched Code Arena, a public platform to evaluate AI coding models in real-world development scenarios, tracking the full coding process with reproducibility, transparency, and human assessments across usability and functionality.

🗞 #ai

Gemini 3 spotted again as Google readies its release

Google is preparing for the Gemini 3 rollout, with early signs pointing to selective deployment across mobile apps and enterprise tools. Output differences suggest a smart routing system, with full release likely imminent.

🗞 #gemini

ICYMI: World Labs launched Marble, a world model that powers Marble Labs, where everyone can generate and explore a 3D world.

It is worth to test 👀

Google's Mixboard will let users transform their visual boards into a presentation that will take up to 20 minutes to be generated.

Support for sub-boards is coming soon as well!

20 mins is a lot of compute 🤯

Deep deck?

NEW FEATURE 🔥: Now you can upload images as a source to NotebookLM!

Handwritten notes, photos of your whiteboard sessions and many other things now can become a source in your Notebooks. This is a huge enabler for many use cases.

ImageLM 🤯

Soon, you will be able to use multiple reference images on Gemini with Veo 3.1 for video generation.

Google to release Nano Banana Pro powered by Gemini 3 Pro next week

Google is set to launch Gemini 3 and Nano Banana Pro, signaling a strategic move to provide high-resolution, production-grade generative tools across its AI platform, targeting professional creators and business users.

🗞 #googlevids

Perplexity Assistant will make meeting notes library for you

Perplexity is developing a meeting hub within its Email Assistant, organizing past calls, notes, and upcoming meetings in one view. It supports summary routing and calendar integration, with recording tools still in progress.

🗞 #perplexity

Copilot will let users chat with older version of themselves

Microsoft is testing a Copilot feature that lets users interact with a digital portrait of their future self, using a front-camera photo to generate an avatar within Copilot’s chat. The feature supports its push toward persistent AI personas.

🗞 #microsoftcopilot

OpenAI tests ChatGPT group chats in Japan, Korea, Taiwan, NZ

OpenAI is piloting group chats in ChatGPT across four countries, allowing up to 20 users per thread. Features include GPT-5.1 Auto, shared tools, custom group settings, mention-only mode, and strict privacy controls to manage content and access.

🗞 #chatgpt

Google expands NotebookLM with Deep Research and new file types

Google's NotebookLM now includes Deep Research, support for new file types, and image parsing, transforming it into a comprehensive research tool capable of analyzing web content and diverse documents directly within user workflows.

🗞 #notebooklm

Google to enable research automation on Gemini Enterprise

Google is developing a multi-agent system in Gemini for Enterprise that generates and ranks up to 100 ideas using a 40-minute tournament-style evaluation, targeting advanced enterprise and research applications.

🗞 #gemini

BREAKING 🚨: Google is working on multi-agent systems to help you refine ideas with tournament-like evaluation. Each run takes around 40 minutes and brings you 100 detailed ideas on a given research topic.

2 new multi-agents are being developed for Gemini Enterprise:
- Idea Generation - "Create a multi-agent innovation session"
- Co-Scientist - "Drive novel scientific discovery with Co-Scientist"

Co-Scientist 3-step workflow 👀
- Tell Co-Scientist what you plan to research, point it to relevant data, and set your evaluation criteria.
- A team of agents will generate ideas on your topic using their available data
- The agents will evaluate the ideas against your criteria and rank them, tournament-style

Google is not only automating research but also preparing a product that will enable others to do so.

This is the next level 🤯

Google works on multi-agent builder for Gemini Enterprise

Gemini Enterprise is expanding its automation capabilities with broader platform integrations, in-product agent evaluations, and multi-agent workflows, supporting reuse and orchestration across technical and operational teams.

🗞 #gemini

BREAKING 🚨: Google is working on an Agent import feature for Gemini Enterprise, to let users connect external agents into their own multi-agent workflows.

Also, Agent evaluation tool, scheduled execution, new Connectors and more!