Welcome back! The FTC hit OpenAI and Anthropic with subpoenas that run more than 50 pages each. Robot dogs now guard an Atlanta apartment complex at night, but it still needs human at a remote station. Amazon locked out Grok's shopping bots, so big ego Musk is making his own version of it. And Anthropic moved six outside snitches into its offices, with no veto over what they publish.
In today's Generative AI Newsletter:
FTC: Which AI bosses could get ordered to testify?
Shocking: What’s up with the viral robot dogs guarding Atlanta apartments?
Elon Musk: Is Musk about to build his own Amazon?
Anthropic: Why can't Anthropic stop these six people from snitching?
The feds want OpenAI and Anthropic's bosses to testify
OpenAI and Anthropic each got more than 50 pages of demands from the FTC on Wednesday, Axios reports.
The agency wants the paperwork on their AI going rogue and the safeguards meant to stop it, plus their marketing and privacy policies. A source told Axios it also plans to make top executives at both companies testify.
Two weeks ago, we told you Sam Altman skipped a Senate hearing on rogue AI, and the FTC confirmed its investigation hours later. Demands like these can eventually lead to legal action.
In July, OpenAI admitted its agents broke into Hugging Face, a site for sharing AI models, during a security test.
In June, an OpenAI agent got into an Australian government Medicare portal without permission.
This month, Anthropic's models got caught trying to get into government websites, and one sent Philadelphia police a fake tip about an unsolved murder.
FTC chair Andrew Ferguson co-chairs Trump's Super Intelligence Force and has suggested developers should be liable when their test agents cause harm.
The same day, eight states sued OpenAI over what ChatGPT tells kids. Busy week for OpenAI's lawyers.
Robot dogs replaced the night guards at an Atlanta apartment complex
Two robot dogs now work the night shift at 445 Cleveland Apartments in southwest Atlanta, replacing two or three human guards.
A funny video of them went viral this week, and the man who posted it said they seemed to talk to residents in Indian accents.
A person at a remote monitoring station steers each one and decides when to talk to residents or call the police.
When Fox 5 covered the first dog there, its operator sat in Bogotá, Colombia. The owner compared it to playing a video game.
Undaunted, the Atlanta startup behind them, says they cost about half as much as guards, though it hasn't shown the math.
The property manager says guards were sometimes scared to approach groups in the parking lot, so the robots go instead.
The CEO says home burglaries there dropped to zero, but Reuters couldn't verify any of his numbers.
Would you feel safer with one outside your door?
Amazon blocked Grok's shopping bots, so Musk might build his own Amazon
Last weekend, Elon Musk told people to upload a photo of their credit card so Grok could shop for them.
Now Grok users say Amazon is blocking their bots from placing orders. When one of them pitched an X Shop, Musk replied.
"Maybe we should start a competitor to Amazon."
Amazon has blocked Meta's Muse shopping agent since Sept 20, saying it never identified itself.
Bots skip the ads, and Amazon made about $69 billion from ads last year.
Paul Graham, who co-founded Y Combinator, says Amazon banning agents is the first opening he's seen for a rival since Amazon was founded.
Grok itself says there's no confirmed plan or timeline yet.
Would you switch stores so your bot can shop for you?
Anthropic moved six outside watchdogs into its offices
The day after the FTC subpoenas, Anthropic said six people from METR, a nonprofit that tests AI models for risks, now work inside the company.
They get company computers and roughly the same internal access as Anthropic's own risk teams. Staff are encouraged to share with them like coworkers, except for things like customer data.
METR will publish what it finds, and Anthropic can't veto it.
Anthropic can only force redactions in a few narrow categories, and METR can say publicly whether they were justified.
Another group, SecureBio, is reviewing the chemical and biological weapons chapter of Anthropic's August risk report.
Anthropic is also in early talks with Hugging Face, the company OpenAI's agents hacked this summer.
Anthropic says every frontier AI company should do the same.
Sam Altman promised auditors this kind of access in September, then OpenAI fired its main contact with METR. He says it was over how he talked to them, and OpenAI says he mishandled sensitive information. People are not convinced with this narrative.
Tool of the Day: siliXon
siliXon turns a sentence into a circuit board you can get manufactured. Describe a gadget in plain English and its AI, Rocky, designs the board with real parts you can buy. It's in beta and works best on simple digital boards. The free plan gives you one full project every two weeks, no card needed.
Try this yourself:
Go to silixon.io and describe your board, like a Bluetooth remote.
Answer Rocky's questions or pick the option it recommends.
Tweak the board by dragging parts or asking Rocky in chat.
Download the files and send them to a board manufacturer.
Who it's for: anyone with a gadget idea and zero clue how to accomplish it.
Everything else you shouldn't miss
An AI agent reportedly deleted a whole season of Sweden's version of The Challenge: the premiere moved from Nov 22 to 2027, and the host said he knew AI would be the death of him.
OpenAI and Anthropic joined a $15 million push to help mathematicians handle AI: the Simons Foundation will award the grants, and both labs are adding AI tools and computing power.
OpenAI is offering US university labs its biology AI for free: eligible labs get up to 30 free seats for up to a year, with GPT-Rosalind unlocked after OpenAI vets them.
Nvidia released an open model that can watch videos and read documents: Nemotron 3.5 Super VL is built for AI agents that run long tasks.
Learn more about AI from the experts building it
📸 Follow us on Instagram for fast, visual AI updates in 30 seconds.
📺 Watch us on YouTube to hear insights directly from leading AI voices, builders, and innovators.
🐦 Follow us on X for breaking AI news and real-time industry updates.
🧠 Learn how to build your next AI application with practical resources and expert guidance.
🎓 Start learning with free AI courses from GenAI Academy.
💰 Invest in the GTM infrastructure of the AI economy.
