Claude can now build live dashboards

Good morning. Claude can now answer a question with a dashboard that keeps itself up to date, and with a short animated video built from the same data. Google also showed off an agent that works around the clock, and three fired OpenAI safety researchers went public with their side.

You’ve put in the work to build an audience. But when the only way to reach your fans is through a social feed determined by algorithms, the platform decides who sees what you make next.

beehiiv gives you the tools to reach your audience directly. On their all-in-one platform, you can turn followers into subscribers, grow your newsletter, and connect it all to your community or podcast. Stop chasing virality on platforms you can’t control. Build a stronger, more sustainable business on beehiiv.

Get started for free.

TODAY IN AI

1. Anthropic launched Claude Dashboards and Claude Motion, both in beta. Dashboards (paid plans) turns a plain-language question into charts that refresh as your data changes. It connects to Redshift, BigQuery, ClickHouse, Databricks and Snowflake, plus CRM tools like Salesforce. Click any number and you can see the query behind it, and each chart shows when it last refreshed. Motion (Team and Enterprise only) makes short animated explainers from your reports and charts. Anthropic says the animations are code, not generated video, so there is no AI footage or AI-made people, and you can edit them or export an MP4. For Enterprise organizations both tools are off by default, and admins switch them on under Organization settings.

2. Google introduced the Gemini agent at Gemini at Work 2026, a single agent meant to handle chat, long tasks, scheduled tasks and event-triggered work. Its work runs in the cloud and keeps going after you close your laptop. It works inside Gmail, Drive, Docs, Sheets, Slides, Chat and Calendar, and also in Microsoft 365 and Slack. Google says it orchestrates across Gemini and Anthropic's Claude models today. The post gives no pricing and no general availability date for the agent itself. Financial Services and Legal versions are in preview, and the customer results quoted, like Bradesco cutting review time from 1 hour to 5 minutes, are self-reported.

3. Three fired OpenAI safety researchers published an open letter disputing the company's account. Jasmine Wang, Tomek Korbak and Mikita Balesni were let go the week before, and OpenAI says they violated policy by "accessing and handling sensitive company information." The three deny mishandling anything and say the firings are "chilling the open culture OpenAI has prized in the past." Wang says she was told she was fired over access to an executive's email that had been delegated to her for recruiting. OpenAI hasn't formally answered the letter. It shared an internal memo saying "We do not terminate employees for raising concerns," and it didn't say which policies were broken.

Top engineers at Anthropic say AI now writes 100% of their code.

Are you using AI to write yours?

These 100+ Claude Code hacks show you exactly how. Sign up for The Code and get:

Claim your free playbook

FROM THE FRONTIER

A sky map built by agents. Anthropic researcher and Johns Hopkins astrophysicist Brice Ménard used Claude Science to build what the write-up calls the first complete ultraviolet map of the sky. Claude coordinated many parallel agents to find public UV surveys, clean and calibrate them, merge them onto one grid, and fill the gaps.

The fine print. About a third of the map, including much of the galactic plane, is predicted rather than measured. In tests where real data was hidden, predictions landed within about 10% of the measurements. An early version also had circular artifacts that passed two rounds of agent review before a human spotted them. The map labels each pixel "measured" or "predicted," and the write-up is Anthropic's own, so treat the framing as theirs.

The reframe. Dashboards that refresh themselves and a sky map that fills its own gaps share a trait: the output looks finished whether or not the inputs were checked. The useful design choices here are the small ones, like a query behind every number and a "predicted" label on every guessed pixel. As AI builds more of what you read, the scarce skill is knowing which parts were measured.

IN THE KNOW

Docs, Slides and Design are out of beta. They're now on every Claude plan, including Free, and Anthropic says 45 million documents, decks and designs have been made in Claude.

Mark your calendar. The standalone Claude Design site stays up until December 14. After that, standalone chats and project comments go away and public project links stop working. Team design systems can be migrated from the Artifacts page.

Admins, note Oct 15. Docs, Slides and Design turn on by default for Enterprise organizations on October 15, though admins can enable them sooner.

Agents with their own email. Google says a Gemini coworker agent can get a persistent role, its own Workspace account and an address on your company domain, and it sees only what is shared with it.

Spend caps for agents. Google's Cloud Billing now supports per-project caps, and when a project hits its cap the agent pauses until someone resumes it with one click. Details are in the same announcement post.

The next A.I. boom could create massive winners just like the 1990s tech surge.

We identified 7 small tech companies positioned to benefit from the next phase of A.I. growth.

See them inside this free report 7 Stocks to Ride The A.I. Megaboom.

View the 7 Stocks

PROMPT STATION

Plan the dashboard before you build it. The failure with a new dashboard isn't the chart. It's tracking numbers nobody acts on. This prompt makes you decide first.

Tip: answer with real decisions from last week, not ideal ones. If a number wouldn't have changed what you did, it doesn't belong on the dashboard.