Welcome, humans.

Okay, so apparently the next AI land rush is not a model. It is the end of your website URL.

ICANN's latest round for new top-level domains has AI companies fighting over endings like .agent, .agi, and .superintelligence (or .si, if you’re Slovenian).

OpenAI apparently applied for 15 strings, Meta applied for 21, and Anthropic wants .anthropic and .claude.

Nothing says "the age of superintelligence is upon us" like a bidding war over what comes after the dot in your web browser…

Here’s what happened in AI today:

• 😺 Anthropic launched Claude Dashboards and Claude Motion

• 📰 OpenAI reported roughly $50B in annualized revenue

• 📰 Biohub partners committed $1.8B to virtual-cell data

• 🍪 Google launched a Gemini work agent

• 💡 Mathematicians are debating how AI changes their job

😺 Claude Can Now Build the Dashboard and the Presentation

So Anthropic just launched two new beta tools that take Claude from answering questions to building something you can actually show people: a live dashboard and an animated explainer.

Here's what's new:

• Claude Dashboards connects to company data, including systems like Salesforce or Snowflake, and builds charts that update as the underlying numbers change. Ask why sales dropped, for example, and Claude can build a live view of the answer.

• Claude Motion turns a report, chart, or idea into a short animation. It generates code instead of AI video footage, so every word, number, and transition stays editable.

• Claude Docs, Slides, and Design are out of beta and available on all Claude plans, including Free.

Finally, you can fix the typo without asking AI to remake the whole thing, be it a doc, slide, new design, or full-on video.

And then Adobe entered the chat: Click Send to Adobe to open a Motion animation directly in Firefly Video Editor. There you can rearrange clips, adjust pacing, color, and audio, add footage, or generate new clips. Adobe's Deepti Pradeep shared a demo of the handoff in action, in case you’re curious.

Adobe's Claude plugin also offers 80+ creative and productivity tools across apps like Photoshop, Premiere, Acrobat, and Adobe Express. The bigger idea: begin the work with AI, then pull in pro editing tools when you need precision.

And yes, this should sound familiar: OpenAI just rolled out GPT-6 Intelligent UI in ChatGPT, which can generate charts, forms, calculators, diagrams, and other interactive interfaces directly inside a conversation.

Anthropic is pushing the same bigger shift from answer → artifact, but with a more work-specific angle: live dashboards tied to company data, then animations that can hand off to Adobe for real editing. Everybody suddenly wants to own the last mile between “AI answered me” and “I can actually use this.”

Dashboards is in beta on paid Claude plans, while Motion is in beta on Team and Enterprise. Claude Docs, Slides, and Design are also now available on Free plans.

Why this matters: AI is moving from providing answers to producing finished work. The real test is whether the dashboard's numbers can be checked and the animation can be corrected without starting over. Pretty charts are easy. Trustworthy ones your boss can actually use? That's the real benchmark.

FROM OUR PARTNERS

Inside Shopify's sandbox for AI shopping agents

Testing AI shopping agents on live stores is messy. Prices change, layouts shift, and bot detection cuts runs short.

Shopify's engineering team built ShopGym to fix that. It turns real online stores into anonymized sandbox shops that reset on demand, then generates realistic shopping tasks to test agents against.

Agents that score well in the sandboxes tend to score well on the live stores they mirror.

The code is open source. See how ShopGym works.

🎓 AI Skill of the Day: Ask for a Decision Dashboard

A dashboard is useful when it helps you decide something. Otherwise, congratulations, you made a prettier spreadsheet.

If you have access to Claude Dashboards, start with the decision you are trying to make, not a list of charts you want. Tell Claude what question matters, what data sources it can use, and which thresholds would change your action.

Then make it show its work. Ask for the definition behind each metric, the query that produced it, and a short explanation of what could make the result misleading. That turns the dashboard from "AI made a number" into something you can actually review.

The goal is fewer mystery charts, more inspectable decisions.

Copy/paste:

Build a live dashboard to help me decide [decision]. Use [data sources]. Start by identifying the 3-5 metrics that would actually change this decision. For every metric, show the definition, underlying query, current value, trend, and the threshold that would change my action. Flag missing data, ambiguous definitions, and anything I should verify before trusting the result.

Have a specific skill you want to learn? Request it here.

UPCOMING EVENT: The Neuron IRL in San Francisco

The Neuron is going IRL on November 18 at 4:30 PM PT! Join us in San Francisco for drinks, bites, special guests, and a live recording of The Neuron podcast. Huge thanks to Slack and Alumni Ventures for making it happen.

It’s free, but space is limited. Save your spot!

📰 Around the Horn

1. OpenAI told investors it reached roughly $50B in annualized revenue at the end of September, below a circulated $68B figure that included partner gross revenue.

2. Anthropic launched a Cyber Mission that gives critical-infrastructure partners its strongest models, on-site engineers, and threat research for grids, water, factories, transport, and government systems.

3. Anthropic updated its Usage Policy to restrict sustained cruelty toward Claude and tighten rules around propaganda, surveillance, weapons, and harmful hardware use.

4. Arena raised a $200M Series B at a $3.1B valuation and launched an Alignment Index for ranking models on unauthorized actions, false attribution, and deceptive completion.

5. Biohub, DOE, NIH, Google DeepMind, Isomorphic Labs, and Meta committed $1.8B to build open, AI-ready biological data for virtual-cell research.

6. OpenAI expanded its education push for teens while independent testing raised questions about age detection and safety safeguards.

🍪 Treats to Try

1. Google's Gemini work agent plans tasks, delegates work, and can route them across agents.

2. Anthropic OSS Scanner scans enrolled open-source projects for vulnerabilities and suggests fixes.

3. Whistle runs a small speech-to-text model locally on a CPU.

4. Durable Actors gives developers persistent actors with their own SQLite state.

5. Rembrandt edits photos locally with RAW support, masks, and on-device AI tools.

6. Pocketty lets iPhone and iPad users manage SSH sessions (secure encrypted remote network access) computer and agent notifications.

7. Atomic turns natural-language engineering processes into verifiable coding-agent workflows with tests, repair loops, evidence, and human approval gates, so you can delegate longer coding jobs without babysitting every step (learn more in our stream w/ Alex from below).

💡 Intelligent Insights

• Scott Aaronson argues the flood of machine-generated math proofs could turn mathematicians into verifiers and explainers of results humans did not discover themselves.

• The Association for Human Mathematics is asking members to keep commercial AI out of research and publication, including a stricter caucus that rejects model use entirely.

• Sasha Rakhlin argues academia should reward better questions, auditing, replication, negative results, and shared datasets as AI takes over more formal puzzle-solving.

• Joachim Klement warns the AI investment boom could end in a 2027 or 2028 market drawdown as data-center capex rises and hyperscaler free cash flow disappears.

• Jono argues DeepSeek 4.1 Flash is close enough to frontier coding quality at dramatically lower cost that Western model economics deserve more scrutiny (particularly the heavily subsidized subscriptions; then again, they won’t be that subsidized if we fix the architecture to be ultra efficient!)

New from The Neuron: AI Explained

What if AI could help turn lifelong genetic diseases into one-time cures? We sat down with Mammoth Biosciences CEO Trevor Martin to learn how his team is using AI to search 34 BILLION proteins, design new CRISPR gene-editing tools, and potentially make some genetic diseases a thing of the past within 5–10 years. This is the kind of AI breakthrough that could actually change millions of lives.

Watch / Listen: YouTube | Spotify | Apple Podcasts

THIS EPISODE WAS BROUGHT TO YOU BY…

Special Shout to Mercor for sponsoring this episode! 

A Cat’s Commentary

Btw: We just launched a robotics newsletter! Sign up for it here.

P.S: Love the newsletter, but only want to get it once per week? Don’t unsubscribe—update your preferences here.