Cowork Dashboards, Kimi K2.6, Gemini in Chrome

 🔥🔥🔥 Three Highly Curated AI Stories

1. Claude Cowork can now build live artifacts

Claude can now build live artifacts inside Cowork, including dashboards and trackers that stay connected to your apps and files. You can open them anytime, they refresh with current data, and everything gets saved in a new Live Artifacts tab with version history.

We built a real-time priorities dashboard in Cowork — pulling from email, Slack, newsletter analytics, and YouTube in one view — and it cost roughly 40% of session tokens to set up. But you only need to build it once, and it's sitting there tomorrow, next week, next month, refreshed and waiting. You describe what you want to track.

Instead of spending fifteen minutes assembling context before every working session, you walk in and the picture is already up to date. The token cost is real and not subtle. But if the output is a reusable dashboard you'd otherwise rebuild or outsource, the math changes pretty quickly.

If you're already using Claude Cowork, this is the week to go build the view you keep wishing you had.

2. Kimi K2.6 is going after serious coding work

An open-source coding model just ran for over 12 continuous hours on a single task — and it's free to use.

Moonshot's Kimi K2.6 is built for the kind of software work that breaks most AI tools: long multi-step projects, not quick autocomplete. You hand it a real engineering task — say, building out a frontend with animated hero sections and WebGL — and it keeps working, making thousands of tool calls, across Rust, Go, Python, whatever the job needs, without stopping to ask you what to do next.

The real story isn't one new model. It's that the baseline for what open models can handle is moving fast. The gap between "closed, expensive, capable" and "open, free, capable" is narrowing in real time.

To be fair: 12 hours of execution on a complex task is impressive on paper, but real-world reliability at that scale is still something you'd want to stress-test yourself before depending on it.

If you care about open-source AI for serious software work, Kimi K2.6 belongs on your radar. Start with Kimi Code on kimi.com.

3. Gemini in Chrome expands to seven new countries

Here's what most people missed about Gemini expanding to seven new countries inside Chrome: the browser is becoming the new AI interface, and most people haven't noticed yet.

The feature itself sounds modest — you can:

• Summarize a long webpage without leaving the tab

• Explain something confusing on the page you’re viewing

• Pull out key points from an article or document

• Help you understand or compare information while browsing

But think about how much of your actual work already lives inside Chrome. Docs, email, research, dashboards, articles. Gemini isn't moving into a new space; it's moving into the space where the work already happens.

Instead of breaking your flow to open a separate AI tool, the assistant is already on the page you're reading. That friction removal sounds small until you realize it's the friction that stops most people from using AI consistently in the first place.

File this alongside Microsoft's Copilot in Edge and Apple Intelligence in Safari. The browser wars are quietly becoming the AI assistant wars. We'll be watching which one people actually reach for.

🔥🔥 Two AI Tools Worth Trying

11. ✨ Gemini memory + chat import

Google is making it easier to switch to Gemini by letting users import both memory and chat history from other AI apps. That means preferences, personal context, and old conversations can move over instead of forcing you to start from zero.

This is especially useful for people who’ve built a lot of context into another AI app and don’t want to retrain a new one from scratch.

2. 🎥 Claude Code workshop

This workshop is from the creator of Claude Code himself, and from what we’ve seen, it will probably teach you more practical vibe-coding in half an hour than a stack of generic tutorial videos. Best for anyone who wants to stop watching Claude Code content and start actually using it well.  

🔥 Things You Didn’t Know You Could Do With AI

Create a cinematic “time freeze” video effect with Seedance 2.0

Seedance 2.0 is a video generation tool, and one fun use case making the rounds right now is the time-freeze effect: one character keeps moving while the whole scene locks in place like a paused movie frame.

1. Start with a reference image so the main character stays visually consistent across the clip.

2. Describe the frozen scene in detail, including the setting, camera movement, lighting, and what gets suspended in mid-air.

3. Add a beat where only the main character moves through the frozen world.

4. End with a reverse effect that restores motion for a dramatic payoff.

Here’s the full prompt!