☀️ TRENDING AI NEWS 🚨 Legal Win: A federal judge ruled the Pentagon's blacklisting of Anthropic as a national security risk was illegal and retaliatory. 🏢 Industry: Nvidia is forecasting $108 billion in quarterly revenue, becoming the first chip company to crack that milestone. 🛠️ Tools: Google Gemini Notebook now lets you import books from Google Play and ask questions, generate summaries, and build AI podcasts from them. 🤖 Robotics: Hugging Face launched the Microduck, a $399 open-source robot you can teach new tricks with reinforcement learning. |
Picture a courtroom in California where a federal judge reads out the words: the Pentagon's designation of Anthropic as a national security supply-chain risk was "illegal and baseless." That ruling landed Thursday - and it's one of the cleaner legal wins an AI lab has scored against the current administration. We've also got AI finding hidden heart disease inside breast scans, Google making your book collection actually useful, and a rollerskating duck robot. Let's get into it.
🤓 AI Trivia
The Pentagon blacklisted Anthropic earlier this year, citing concerns about the company as a 'supply chain risk.' But what was the specific trigger that reportedly caused the administration to retaliate against Anthropic in the first place?
🔢 Anthropic published research claiming Claude outperformed GPT-4 on military benchmarks
🔢 Anthropic set 'red lines' on unacceptable military uses of its AI and made them public
🔢 Anthropic refused to sign a federal AI procurement contract
🔢 Anthropic's CEO testified against the administration in a Senate hearing
The answer is hiding near the bottom of today's newsletter... keep scrolling. 👇
| ⚠️ A Federal Court Just Ruled the Pentagon Broke the Law - Over an AI Lab | |
The Trump administration punished Anthropic for drawing ethical lines
In February, the Trump administration labeled Anthropic a national security supply-chain risk - a designation that, if it had stuck, could have cost the company billions in lost government contracts and serious reputational damage. On Thursday, a federal judge in California ruled that designation was unconstitutional and retaliatory, finding the government had punished Anthropic specifically for setting public limits on how its Claude models could be used in military contexts.
The lawsuit, filed in March, argued the government had singled Anthropic out for criticism rather than any genuine security risk. The judge agreed - calling the designation "illegal and baseless." It's a significant moment for AI labs navigating an increasingly politicized regulatory environment, where speaking publicly about safety limits can apparently trigger federal penalties.
The bottom line
This ruling matters well beyond Anthropic - it signals that AI companies can't be quietly punished through procurement blacklists for taking public ethical stances, and it sets a precedent other labs will point to if they face similar pressure. Follow the ongoing coverage of AI regulation and legal battles in the AI space at Daily Inference.
| 🔬 AI Spots Heart Disease Hiding in Routine Mammograms | |
One scan, two life-saving diagnoses at once
Here's one that quietly blew my mind. Researchers have found a way to use routine mammograms - already being done for breast cancer screening - to simultaneously identify women at risk of cardiovascular disease, the world's leading cause of death in women and one of its most underdiagnosed. The key is AI analysis of the scan images.
The AI system analyzes calcification patterns in breast arteries visible on mammograms - patterns that correlate strongly with cardiovascular risk. Women already showing up for cancer screening could, in the same appointment, get a flag that sends them to a cardiologist. No second scan, no extra cost, no extra visit.
One scan doing double duty
Cardiovascular disease kills more women each year than any other condition, yet it's routinely missed or dismissed in clinical settings. The fact that an existing screening infrastructure could be repurposed to catch it earlier - using AI to read data that was always there but ignored - is the kind of practical, unglamorous breakthrough that actually saves lives.
The bottom line
If this moves from research into clinical practice, millions of women who get annual mammograms could receive an early cardiovascular warning at no extra cost - which is exactly the kind of AI application that makes skeptics pause.
| 🛠️ Google Gemini Notebook Can Now Read Your Books | |
Your digital bookshelf just became a research assistant
Google's AI note-taking app, Gemini Notebook, has added a feature called "Expert Intelligence" that lets you pull books directly from Google Play Books into your workspace. Once imported, you can ask questions about the material, generate study plans, build infographics, or even spin up an AI podcast based on the book's contents.
This is a natural extension of what made Notebook useful for research documents and PDFs, but now applied to commercially published books. For students, researchers, or anyone using Google Gemini as a reading companion, the ability to interrogate a full book rather than skim it is genuinely different from anything currently baked into a mainstream reading app.
The bottom line
If you've got a library of Google Play Books collecting digital dust, this is worth revisiting - being able to ask a non-fiction book specific questions and get cited answers is a legitimately useful upgrade to how most people interact with books they've already purchased.
| 🤖 Hugging Face's New Robot Costs $399 and Rollerskates | |
Open-source robotics for everyone, now duck-shaped
On the open-source front, Hugging Face's Pollen Robotics has launched the Microduck - a one-eyed, ten-inch-tall biped robot available for $399 in cream, graphite, lavender, and sky blue. It can pick up socks and markers, kick around a ball, and yes, rollerskate. Preorders are open now, with shipping planned before Christmas 2026.
What makes this more than a novelty is the open-source angle. CEO Clem Delangue describes it as a robot you can teach new tricks using reinforcement learning - meaning developers and hobbyists can actually train it to do new things, not just run preset routines. At $399, it's priced in reach of individual researchers and educators, not just labs with procurement budgets.
Speaking of building things fast - if you're inspired to spin up a quick project page or landing site around something like the Microduck, 60sec.site is an AI website builder that gets you from idea to live site in under a minute. Worth bookmarking.
The bottom line
The Microduck isn't going to replace industrial robotics anytime soon, but a sub-$400 trainable open-source robot is exactly the kind of hardware that accelerates how quickly the broader community experiments with and improves robot learning.
| 🎤 80 Actors Just Signed an Open Letter Against AI Voice Cloning | |
Nicola Coughlan and Hugh Bonneville want legislation now
About 80 actors and performers - including Nicola Coughlan, Hugh Bonneville, Matt Lucas, and Luke Evans - have signed an open letter addressed to Manchester Mayor Andy Burnham calling for legislation to protect voice ownership. The campaign, called Save Our Voices Now, targets the practice where AI technology replicates a performer's voice without consent or compensation.
The letter specifically calls for legal protections that would require consent and payment before any AI can clone a professional voice. This is part of a broader wave of entertainment industry pushback against AI - similar battles have played out in Hollywood over likeness rights and in the music industry over AI-generated vocals. The performers are framing this as a labor rights issue, not just a tech concern.
The bottom line
Voice cloning is moving faster than the law - and when A-list names attach to a campaign, it usually accelerates the legislative timeline. This one is worth watching if you work anywhere near audio production, voiceover, or AI-generated media.
| 🌎 Trivia Reveal | |
The answer is B - Anthropic set 'red lines' on unacceptable military uses of its AI and made them public. According to the lawsuit and the court's ruling, the Trump administration's Pentagon retaliated against Anthropic specifically because the company had publicly defined limits on how Claude could be used in military contexts. The judge found that punishing a company for taking a public ethical stance was illegal - and the blacklisting was overturned.
| 💬 Quick Question | |
The Anthropic ruling raises a big question: should AI labs be legally protected when they publicly set ethical limits on how their models get used - even by governments? Hit reply and tell me where you stand. Do companies have a right to say no to military use cases, or does that create its own risks? I read every response.
That's it for today - enjoy the rest of your Saturday. We'll be back tomorrow with more from the world of AI. And if you want to explore any of today's topics deeper, head over to dailyinference.com for the full archive.