In partnership with
☀️ TRENDING AI NEWS 🤖 Anthropic: Warns it is very close to recursive self-improvement, where AI trains itself and capabilities accelerate sharply. 🏢 OpenAI: Pilots a Grid Protectors Program in Effingham County, Georgia, promising up to 1,000 megawatts back to the grid in emergencies. 🛠️ Apple Watch: New Audio Intelligence features that rewind and recap conversations are now in beta. 🚨 NYT vs Brave: The New York Times and other outlets sue Brave over alleged AI scraping and paywall bypassing. |
Imagine an AI lab telling you, in public, that the thing the doomers have been warning about is nearly here. That happened yesterday. And it was not even the strangest AI story of the last 48 hours.
| 🛠️ Tool of the day | |
NotebookLM - Google's AI research assistant. You add documents, web pages or videos, then ask questions and get answers with citations to your own sources. It also builds study guides and audio overviews, and it helps students, analysts and anyone who reads a lot. It has a free plan.
| ⚡ Quick hits | |
Meituan released LongCat-2.5, a multimodal model with 1.6 trillion total parameters and about 48 billion active per token, built for long coding tasks. Read more →
SemiAnalysis launched GDPx, a benchmark that spans domains from chip design to cybersecurity, and says even the best model scores just 12%. Read more →
| 🤓 AI Trivia | |
What does the acronym RSI stand for in AI safety discussions?
The answer is hiding near the bottom of today's newsletter... keep scrolling. 👇
| ⚠️ Anthropic Says It Is Very Close to Self-Building AI | |
The point where AI trains itself
Anthropic warned that it is very close to RSI capability, the point at which AI gets so good at training itself that it spawns a dramatic acceleration of AI capabilities.
To the AI doomers, RSI represents the cataclysm from which there is no return. Hearing a leading lab say it is nearing that threshold, rather than a critic speculating about it, is a notable shift in tone for the industry.
The bottom line
Keep an eye on how other labs respond. If rivals echo the claim, the safety conversation moves from theory to a near-term planning problem.
Two days of free email tactics, starting Nov 12
GURU kicks off Nov 12. Join the 29,000+ marketers who showed up last year, 100% free and virtual.
GURU Conference, powered by Constant Contact, is back November 12–13. It's 100% virtual and 100% free, and it's packed with tactics you can use right away on newsletters, deliverability, design trends, AI, and what NOT to do with email.
What to expect:
Speakers including Molly Ringwald, Dan Levy, Amy Porterfield, Frank Vella + more
Sessions from world-class marketers at top brands
Networking every day
DJs, world-record breaking, and an alarming amount of rom-com fun!
Last year, 29,000+ marketers joined us. This year is even bigger, and spots are limited.
| 🏢 OpenAI Pitches Data Centers as Grid Protectors | |
Power plants with a PR mission
OpenAI announced it will pilot its Grid Protectors Program at an upcoming data center in Effingham County, Georgia. The company promises to free up to 1,000 megawatts of power during extreme weather or emergencies, enough to power air conditioners in around 700,000 homes.
The timing is no accident. Opposition to data centers is spreading across the US as communities worry about the strain on local power and prices. Elsewhere, TechCrunch reports OpenAI is even weighing whether to rebrand data centers as 'AI compute campuses'.
The bottom line
This is as much a local-politics move as an engineering one. Whether a promise to give power back wins over skeptical neighbors is the real test.
| ⚖️ A Prosecutor, an Imaginary Case, and the High Court | |
AI hallucinations reach the justice system
A senior Crown Prosecution Service barrister cited a nonexistent legal precedent in a high court case after an AI search, in an alleged child rape case. The Guardian reports the error adds to dozens of AI-related mistakes plaguing court proceedings.
Hallucinations are a known failure mode of language models, but the stakes here are very different from a bad restaurant recommendation. The story has sparked fresh concerns about fabricated citations seeping into the legal system.
The bottom line
If you use AI for research in any high-stakes job, verify every citation against a primary source. The tool is fast, but it does not carry the professional liability you do.
| 🛠️ Apple Watch Learns to Rewind Conversations | |
Audio Intelligence enters beta
Apple has started beta testing new Audio Intelligence features on Apple Watch that help you remember what someone said or take notes on conversations. Think of it as a recap button for real-life talk.
On the business side, Apple hinted on its Monday earnings call at ways to spur revenue through third-party AI app 'opportunities', without saying exactly what it has in mind.
The bottom line
A wrist-based recap tool is a genuinely practical use of on-device AI. Watch for how Apple handles privacy when recording the people around you.
| 🚨 The New York Times Sues Brave Over AI Scraping | |
Scraping and paywall bypass allegations
The New York Times and a group of other publications have filed lawsuits against the company behind the Brave browser. The complaint filed by the New York Daily News, the Chicago Tribune and the Orlando Sentinel alleges Brave Software monetized an index of scraped websites and let users bypass paywalls.
TechCrunch reports the suit also alleges the scraped content was distributed to customers like Anthropic and Perplexity. It is the latest front in the ongoing fight over AI copyright.
The bottom line
Publishers are now going after the data pipeline, not just the model makers. Anyone downstream of a search index could end up in the crosshairs.
| 💸 Garry Tan's Token Bill Hits $200K a Year | |
Life with an army of AI agents
Y Combinator's Garry Tan has jokingly admitted he has suffered from 'cyber psychosis' and sometimes barely sleeps because he is so excited to work with his AI agents. On Walter Isaacson's podcast The Builders, he put a price tag on the passion: up to $200K a year on tokens.
If you want to see what your own agent habit might cost, try our Token Calculator. And if you just need a quick site up without burning tokens, our sponsor 60sec.site is an AI website builder that gets you live in about a minute.
The bottom line
Token spend is becoming a line item that rivals salaries for power users. Expect more conversations about what that spending actually ships.
| 🌎 Trivia Reveal | |
The answer is Recursive Self-Improvement! It describes AI systems that get good enough at training themselves that capability gains compound, which is exactly what Anthropic says it is very close to.
| 💬 Quick Question | |
How much do you spend on AI tools or tokens each month, and is it worth it? Hit reply and tell me - I read every response!
That's it for today. Catch the next edition at dailyinference.com - see you tomorrow!
| 🎁 Share Daily Inference | |
Know someone who wants AI news in 5 minutes a day? Share your link. 1 referral unlocks the AI Tools Starter Kit. 5 referrals unlock the AI Insider Briefing.