Claude Joins Your Meetings, Google Launches Gemini 4 Hydrogen & Anthropic Files For Its IPO

Today's AI news: Anthropic launches Claude Meetings, which joins your Google Meet calls and builds docs, whiteboards and plans your team can edit while you're still talking. Google DeepMind releases Gemini 4 Hydrogen, its fast everyday model, which edges out the bigger Gemini 4 Argon on the Artificial Analysis Intelligence Index. Anthropic files to go public on Nasdaq. MBZUAI's Institute of Foundation Models releases PAN 2, an open world model that predicts what happens next. And Google's Gemini smart glasses, made with Samsung, Warby Parker and Gentle Monster, go on pre-order. Chapters: 0:00 Today's AI News 0:23 Claude Meetings 1:14 Gemini 4 Hydrogen 1:55 Anthropic IPO 2:34 PAN 2 World Model 3:20 Google Smart Glasses Blend Roundup 2026-11-12 https://x.com/claudeai/status/2120973994191094179 https://x.com/GoogleDeepMind/status/2120940338869055788 https://x.com/AnthropicAI/status/2120841731302154708 https://x.com/IFM_AI/status/2120199814696828942 https://x.com/Android/status/2120885368417333468 Claude starts turning up to your meetings and doing the homework before they've even finished. Google ships Gemini 4 Hydrogen, a fast model that beats its own bigger sibling, Anthropic files to go public, An open world model predicts what happens next, And Google's smart glasses finally go on pre-order. Here's today's AI news. Anthropic has launched Claude Meetings. You add Claude to a Google Meet call, and it joins as a visible participant. Ask it something out loud or in the chat, and it answers. But the real point is that it builds things while you talk: docs, whiteboards and plans that your teammates can edit before the call has even ended. Afterwards, every meeting gets its own page in Claude, with notes, action items, the transcript, and whatever got made along the way. It doesn't save audio, and by default it won't join calls with outside guests. It's rolling out in beta on Team plans, where it's on by default, and Enterprise, where admins switch it on. Artifacts made in calls are free until February, and Zoom and Teams support is coming soon. Google DeepMind has released Gemini 4 Hydrogen, the fast, everyday model in the Gemini 4 line. It effectively replaces 3.8 Flash, and Google says it finishes tasks two and a half times faster. The surprise is that it beats Gemini 4 Argon, the bigger frontier model, on the Artificial Analysis Intelligence Index, scoring 54 at max settings to Argon's 50. It has a million-token context window, and costs a dollar fifty per million input tokens and seven fifty for output. It's out now in the Gemini app, AI Studio, the API and Google Search. Anthropic has filed to go public on Nasdaq, under the ticker A-N-T. The filing sets a range of seventy to seventy-eight dollars a share for 825 million shares, which would value the company at around a trillion dollars. The numbers behind it: revenue of 4.6 billion dollars last year, 33 billion in the first nine months of this one, and a run rate of about 81 billion as of October. It stays a public benefit corporation, with a Long-Term Benefit Trust electing most of the board, and the co-founders keeping majority voting control. Goldman Sachs and Morgan Stanley lead the deal, and final pricing hasn't been set yet. The Institute of Foundation Models at the Mohamed bin Zayed University of Artificial Intelligence has released PAN 2, an open world model. A language model predicts the next word. PAN 2 predicts the next world. You describe a scene and an action, and it generates video of what happens next, grounded in time and physics, for up to twenty seconds at 24 frames a second. Under the hood, a language model writes out the next state of the world, and a 29-billion-parameter diffusion model renders it. The weights and training code are open under Apache 2.0. IFM is candid about the limits though: in roughly two of every three long plans, at least one step doesn't happen. Google's smart glasses are finally on pre-order. They're built with Samsung and come in frames from Warby Parker, starting at 395 dollars, and Gentle Monster, from 449. There's no display. Instead, you get Gemini in your ear through open-ear speakers, plus a 12-megapixel ultra-wide camera. You can ask what you're looking at, get directions, take calls, or translate conversations across more than 70 languages. Battery life is up to nine hours, and they work with recent Pixels, Galaxy phones and iPhones. Sales start December the tenth in the US, and Gentle Monster frames are also coming to South Korea.