Hello AI Enthusiast,
This week AI got a lot more practical. Claude can now build dashboards that update themselves and turn them into short animated videos, Google's new Gemini agent gets its own work email, and Google also released a Mac note-taker that keeps your meetings on your device. We also have GPT-6 for everyone, a much cheaper Claude Haiku and new models from Meta. Let's get into it.
The Big Picture 🔊
Claude now builds live dashboards and animated videos
Anthropic launched Claude Dashboards and Claude Motion. Dashboards connect to sources like Snowflake, BigQuery and Salesforce, stay current as the data changes, and show the query behind any number you click. Motion turns a prompt or a dashboard into an animated explainer you can edit and download as an MP4. Dashboards are in beta on paid plans, Motion on Team and Enterprise.
|
||
|
Dashboards were already one of the most common things people built with artifacts. The difference now is that they stay live, and you can check where every number comes from. |
Google's Gemini agent gets its own email address
Google announced the Gemini agent, one agent for chat, long tasks and code that keeps running in the cloud after you close your laptop. It works across Workspace, Microsoft 365 and Slack, and can use Gemini and Claude models. Coworker agents get their own Workspace account, with an email, calendar and Drive, and only see what you share. It's in private preview, with no date yet for a wider release.
|
||
|
The agent gets its own address in your company directory, like a new colleague, so it's clear what it can reach and what it's for. If your team works in Google Workspace, this could mean one AI workspace instead of two. |
Google releases an on-device meeting note-taker for Mac
Google released AI Edge Foresight, a Mac app that listens to your online or in-person meetings while you jot quick notes, then turns them into full notes using the transcript. Audio and transcript are processed on your device. You can connect project folders, Google Drive and your calendar, and chat with your notes. Google built it to show off EmbeddingGemma 2, its new small model.
|
||
|
Documents show what a team decided. Meetings show how and why, and most of that is lost unless someone records it. A note-taker that runs on your laptop makes recording easy and keeps your data with you. |
All of this only helps if your team knows what to reach for, and when. That is the gap we close. We train companies to use tools like these with confidence, starting from how people actually work. Want to see what that looks like for your team?
Bits and Bobs 🗞️
OpenAI made GPT-6 available to everyone in ChatGPT, Free and Go included. Its new Intelligent UI builds answers with charts, buttons, forms and small tools when a question needs them.
Anthropic released Claude Haiku 5.5, its fastest and cheapest model, about 75% cheaper to run than Haiku 4.5 on average. It's built for high-volume work like summaries, classification and live support.
Meta introduced Muse Wonder and Muse Wonder Full, its most capable models for personal and work agents, built around memory and staying on task. Muse Wonder is available through the Meta Model API and Muse Code, and will soon power the Muse app.
Google brought Gemini 4 Argon to the Gemini app for AI Pro and Ultra subscribers, with a coding interface in the Mac app and connections to more than 40 apps, including Adobe and Shopify.
Anthropic updated its Usage Policy, effective November 12. New rules cover deceptive campaigns, surveillance and autonomous hardware, and the blanket ban on personalized campaign targeting is gone.
Anthropic shared how it fights distillation attacks, where others copy Claude's reasoning to train their own models. The reasoning extracted fell 10 to 100 times, and tens of thousands of accounts were banned.
Anthropic's Frontier Red Team found that small AI models may soon be able to copy themselves across cheap devices like TV boxes, with no data center needed. If small models keep improving at the same pace, they could get there by the end of 2026.
OpenAI disrupted two influence operations from Russia and Iran that used ChatGPT to write fake articles and reports. One placed almost 100 articles in about a dozen outlets under seven fake bylines.
Google launched Playground, an experimental platform where you create and share games with text prompts, no coding needed. It's US only, for users 18 and older.
From Our Founder’s Channels 🤳
In his recent LinkedIn post, Gianluca Mauro argues that an engineering approach to instructional design is key to using AI meaningfully at scale. Read it here.
That's a wrap on our newsletter!
If you're looking to build AI skills in your team, check out our Corporate Training. Every program is tailored to your team and built around hands-on practice.
Catch you next week! 👋
