- AI Valley
- Posts
- AI Labs Prepare for Disaster
Together with
howdy, it’s Barsee again.
happy friday, AI family, and welcome back to AI Valley.
here are the biggest things worth knowing today:
Claude can now build live dashboards and animated videos
AI companies are quietly preparing for a major AI disaster
Three fired OpenAI researchers are pushing back
Plus trending AI tools, posts, and resources
Let’s dive into the Valley of AI…
SPONSORED
Courtesy: Republic
On September 25, in front of a live audience in downtown Los Angeles, The Gifted won the Future Vision XPRIZE, the world's largest sci-fi competition to imagine more positive visions of the future.
The story: an eleven-year-old genius rebuilds his dead mother as an AI. She starts guiding strangers to save lives, and the government decides her kindness is the greatest threat it has ever faced.
Now it's moving toward production with Range Media Partners, the studio behind A Complete Unknown and Longlegs.
You don't have to just watch it happen. The film is open to investors on Republic, giving you a chance to back the award-winning project before it reaches the screen.
THROUGH THE VALLEY
1/ Claude can now build live dashboards and animated videos. Two new things: Claude Dashboards and Claude Motion.
Dashboards connects to company data (Salesforce, Snowflake and the rest) and builds charts that actually update when the numbers change.
Motion takes reports, charts or just ideas and turns them into animated explainers. The clever bit is it uses code instead of AI video, so the text, numbers and animations stay editable. There’s also a Send to Adobe option so you can keep going in Firefly Video Editor. Adobe’s Claude plugin already brings 80+ tools from Photoshop, Premiere, Acrobat etc into Claude, so the two are clearly getting friendlier.
Dashboards is in beta for paid users. Motion is Team and Enterprise only for now. Claude Docs, Slides and Design just came out of beta and are available on Free plans too.
Interesting timing with OpenAI’s Intelligent UI launch. Both companies are pushing past pure text, but Anthropic still feels more focused on the tools people actually use day-to-day. I’ll be trying Dashboards as soon as I get access; live updating charts from real company data is something I’ve wanted for a while.
2/ AI companies are quietly preparing for a major AI disaster. Axios reported that executives at OpenAI, Anthropic and other labs are privately war-gaming a public and political backlash after a catastrophic incident — possibly as soon as 2027. Some insiders apparently expect something within the next 6–12 months. The main fear is an AI-powered cyberattack that knocks out banks, internet, electricity or water.
OpenAI confirmed they run preparedness exercises but stressed these scenarios aren’t inevitable. Anthropic declined to comment.
Worth keeping in perspective: these are scenarios being discussed, not evidence that something is about to happen. Still, the fact that the people building the stuff are running these drills says something.
3/ Three fired OpenAI researchers are pushing back. Tomek Korbak, Jasmine Wang and Mikita Balesni were let go for allegedly mishandling sensitive information. They’ve published an open letter disputing the accusations and saying OpenAI’s shifting internal rules are making people afraid to speak up.
They claim their actions were within policy (including outside research and access to an executive’s email for recruiting). Wang says she accidentally opened a sensitive email and reported it within minutes. They also deny leaking anything about models becoming harder to monitor.
OpenAI told TechCrunch the firings involved a “pattern of misconduct” and weren’t retaliation for raising safety concerns. The researchers are also worried the firings could undermine plans to bring independent auditors inside the company. OpenAI says it agrees with their recommendations, but whether those plans actually move forward is another question.
TRENDING TOOLS
Claude Dashboards and Claude Motion - ask Claude to turn your data into live dashboards and your ideas into animated explainers
Interface - the primary hardware for the agent era. It lets you control all your agents from your hand
REA - one MCP for reverse engineering across applications and runtime behavior
Voyager - the open harness for S-tier video, graphics, and games
Odyssey-3 - most powerful foundation world model yet. A new state-of-the-art in Physics-IQ, and powers robots, trains AIs, and generates interactive experiences
Anthropic OSS Scanner scans enrolled open-source projects for vulnerabilities and suggests fixes
Fo - a personal AI that hires humans for the tasks AI can’t do yet
GoodVibes - filter X, Reddit and YouTube by vibes, in plain English. Runs on Jev
WHAT I'M CONSUMING
We reverse-engineered ChatGPT Intelligent UI. Here's how it actually works
The state of AI report 2026
What's new in Claude: September recap
THE VALLEY GEMS
THAT’S ALL FOR TODAY
Thank you for reading today’s edition. That’s all for today’s issue.
💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee
👍️ New reader? Subscribe here
Thanks for being here.
REACH 100K+ READERS
Acquire new customers and drive revenue by partnering with us
Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.
If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.