Read Online | Sign Up | Advertise | | | Good morning, C, and welcome to our 11,494 new readers who joined yesterday. Frontier labs keep saying that the money AI makes should be shared with the public, but so far that’s mostly lived in research and warnings. Anthropic just put real stock behind it. | Its new Public Benefit Fund grows as Anthropic’s value rises, and the company says it could even send money straight to groups like workers hit by AI. With public trust in AI sliding, cash in people’s hands might do more for the industry’s image than another essay ever could. | | In today’s AI rundown: | Anthropic designs a pay package for the public The Rundown Roundtable: Our AI use cases Turn a content spreadsheet into a working app Anthropic data shows junior employees losing ground
|
|
| | | | ANTHROPIC | |  | Image source: Anthropic |
| The Rundown: Anthropic just introduced the Anthropic Public Benefit Fund, a limited liability company (LLC) that would get 5% of the company’s equity if shareholders sign off (about $48B at its May valuation), plus more as Anthropic’s value climbs. | The details: | As a company-owned LLC, Anthropic said the fund could back nonprofits or “direct distributions to specific populations,” with no stock going to lobbying or politics. Each 50% jump in Anthropic’s value past an unannounced baseline could add 2% at the board’s call, its answer to tech’s growing executive pay scales. The company identified AI’s toll on workers as one target, building on $200M it committed in June to study job disruption and test policy responses. Anthropic now counts about 20% of its shares as set for public good, between the new fund, the co-founders’ 80% pledge, and stock it matches when staff donates.
| Why it matters: Universal basic income and other public-benefit structures have been floated in the AI age, and Anthropic’s fund and OpenAI’s foundation are the clearest signs yet of the labs putting real equity behind that. But until money actually reaches anyone, it will be hard for the public to separate these moves from a convenient PR play. |
|
| | TOGETHER WITH STACK AI | | | The Rundown: With the detail of Palantir and the scale of Salesforce, StackAI is the enterprise AI agent platform with a dedicated team of AI experts who handle integration, security and on-prem deployment, and get your agents into production on your data. | With StackAI, you get: | A no-code platform for complex use cases An AI strategist who tracks your ROI An engineer who builds your custom workflows
| Meet your AI deployment team. |
|
| | THE RUNDOWN ROUNDTABLE | | | The Rundown: The Rundown Roundtable is a weekly feature where we poll members of The Rundown staff about how we use AI in our work and daily lives. | Brooke, Community Engagement Lead: While traveling abroad, my partner and I became quite sick near the end of our trip and remained ill after returning home. Because we were in a country where we did not speak the language, we used Google Translate at the pharmacy and used Claude to track our symptoms as they changed from day to day. | Claude helped us keep the symptom history organized and research what to ask for at the pharmacy, including differences between the medicines we were familiar with and those commonly used in that country. When our symptoms worsened after we got home, we continued using Claude as a symptom tracker and photographed everything in our medicine cabinet. We often forget what we bought or why, and medicines in the United States can be marketed for different uses in ways that are not always clear. | Using the symptom history and medicine photos, we asked Claude what we might be able to use, what to avoid, when to take things, and when to contact a doctor, use telehealth, visit urgent care, or go to the emergency room. It also surfaced newer research suggesting that some medicines once thought to be effective may no longer be supported by current evidence. | This simple workflow helped us manage the situation and decide whether we needed to buy anything new or seek additional care. |
|
| | AI TRAINING | | | The Rundown: In this guide, you’ll use Airtable Omni to turn a content plan into an app that shows what needs review, who owns it, and what’s overdue. Then you’ll add a checkbox that moves a complete draft into review. | Step-by-step: | Create an empty practice base in Airtable and open a new Omni chat. Check access to the upgraded version and turn off Internet access for the exercise Upload content plan and campaign list. Ask Omni to propose the tables and campaign links before building them. Keep missing owners and dates blank Verify the links, then build a review page with counts, a content list, and owner filters. Check the results and change a date to test the page, then restore it Ask Omni to create an automation that moves a complete draft into review when you tick a checkbox. Inspect the setup and test one record before turning it on Try an item with a missing owner and one that is approved. Neither should move. Test a second valid draft to check that the automation updates the right item
| Pro tip: A missing owner should stay blank in the data. Writing “Unassigned” into that field could make an incomplete draft pass the check. Pro members get the practice files, prompts, and answer key. |
|
| | PRESENTED BY COLLIBRA | | | The Rundown: Most orgs can’t list every model, agent and endpoint they run in production, or say who owns each one and what data it uses. Collibra’s AI Trust Score Assessment shows exactly where your AI governance gaps are. | The assessment includes: | 12 questions across six areas, from model inventory to traceability A maturity level plus a 90-day plan to close your gaps Answers that stay in your browser until you request a report
| Run the assessment. |
|
| | AI RESEARCH | |  | Image source: Anthropic |
| The Rundown: Anthropic published a study on how AI has changed its own workforce, finding that employees’ agents logged 6x the hours of the people running them by August, while junior roles make up a shrinking share of the company’s technical hiring. | The details: | Juniors went from 27% of Anthropic’s technical hires to just 9%, and about a third of its machine learning managers said AI had them hiring less. Engineers push 5x the code changes they did six months back, with Anthropic finding agent code is deleted or rolled back no more often than human code was. Claude agents sit idle about 80% of the time in back-and-forth work sessions while they wait on a person to respond, leaving humans as the bottleneck. Year-plus staffers switch departments 50% more often than in 2025, and Anthropic is building AI-risk alerts per role to move people to less exposed jobs.
| Why it matters: The hiring numbers inside one of AI’s top labs line up with what CEO Dario Amodei has been warning about for over a year, but the other big reveal is how much of the job is now keeping up with Claude. Agents are producing more than staff can read, which might also be why labs are racing toward autonomous AI researchers. |
|
| | | | COMMUNITY AI WORKFLOW OF THE DAY | ▸ Edwin built an AI coding playbook that shipped a 3D game | Today’s workflow comes from reader Edwin Vermeer: | “I’ve been experimenting with agentic coding for quite some time, and I’ve noticed that the tasks I let the agent handle keep getting bigger. For larger tasks and projects, I gradually moved to a standard workflow. | I use this workflow professionally and in my side projects. My most recent side project is Quadrim, a four-in-a-row game played in a 4×4×4 cube. I made it open source and included all the prompts for both the code and the artwork. It took me about eight working days to publish the first version to the app stores. | During the pre-implementation steps, I always add this instruction to my prompt: ‘Ask clarifying questions until any ambiguities are resolved.’ Extra fine-tuning of the idea, specifications, or plan pays off during implementation. When I let the agent build code, I ask it to use TDD, perform a code review at the end, and resolve the major and critical findings.” | See the workflow on our Community Workflow Hub → | | What have you fixed or built with AI lately? Share your workflow, and you could be featured right here in The Rundown. |
|
| | | 📝 Goodnotes - Write by hand, lose nothing. Searchable handwriting that behaves like text and connects to ChatGPT and Claude* 🧠 Muse Spark Open - Meta’s open-weight AI for coding and long-horizon agents 🎧 Instant Soundtracks - ElevenLabs’ tool to add scores and effects to clips 🗣️ GPT-Live-1.1 - OpenAI’s new voice AI with more natural multilingual accents
| *Sponsored Listing |
|
| | | AI coding tool OpenCode said hackers may have grabbed the OpenAI and Anthropic account keys some users saved, tying the break-in to a mass effort to copy AI models. | Krea released Krea 3, a new image editor with a focus on fast, precise changes and consistency, claiming 3x the speed of Google’s Nano Banana Pro at under half the cost. | Google signed Axios, CNN, and the Financial Times to its paid AI news program, which now tops 400 outlets, as AI answers keep cutting into publishers’ search traffic. | Odyssey posted a demo wiring a simulated fruit fly’s brain into its Odyssey-3 model that generates virtual worlds, letting the digital bug fly through them with no human steering. | Chinese AI lab DeepSeek updated its V4.1-Flash model, which it says now finishes tasks faster with shorter answers, and opened its in-house web search tools to devs. |
|
| | | | That's it for today!Before you go we’d love to know what you thought of today's newsletter to help us improve The Rundown experience for you. | | See you soon, | Rowan, Zach, Shubham, Jennifer, and Nate — the humans behind The Rundown | |
|
| |
|
