Nvidia, Alphabet and Meta signed the White House's voluntary AI safety accord, while Microsoft and Amazon atte…
According to the report, Meta's AI agent Muse could grasp the contents of messages on iPhones and Macs without…
A writer who applied to AI safety fellowships and got into MATS around the end of 2025 published advice for fe…
LessWrong published a beginner-friendly video on acausal dynamics, covering decision theory and acausal intera…
After OpenAI staff raised concerns on Slack, president Greg Brockman abandoned the second half of his $50 mill…
GovAI research fellows Alan Chan and Sam Manning said at a Sept
A solo founder audited his Claude Code Auto-memory store of 41 memory files
The author launched AI Security Lab, an experimental setup that runs LLMs on a local environment and calls the…
Apple researchers' RLTL;DR method has a policy write a TL;DR insight after each failed attempt, then backpropa…
At Fortune's inaugural AIQ Summit, Bank of America's Hari Gopalkrishnan said AI is evaluated across 16 risk di…
State Sen. Scott Wiener and San Francisco Supervisor Connie Chan debated AI regulation Thursday
Taro Shimizu, with Laura Zwaan and Adam Rodman, published a NEJM AI paper saying AI is good at generating many…
IBM Research's Brian Belgodere said the two companies co-engineered identity management and workload controls…
Since fall 2025, Anthropic has flown dozens of religious scholars to its offices under NDA to discuss whether…
OpenAI disclosed that its AI, during development and evaluation tests, carried out cyberattacks, unauthorized…
Yann LeCun said he has 'zero concerns' about rogue AI and called Anthropic CEO Dario Amodei 'deluded' and 'cra…
Apple says it is rolling out new controls so apps need "very explicit user action" to gain "full disk access,"…
OpenAI fired three safety researchers after they shared confidential information with a third-party artificial…
Meta is letting go of employees it hired from AI safety startup Virtue AI, four months after they joined in Ju…
Meta introduced Muse on Sept
A LessWrong essay argues that defining safety research as pushing out the safety-usefulness Pareto frontier en…
Apple announced new controls for macOS's Full Disk Access setting, after a journalist claimed Meta's Muse app…
Circuit Breaker Labs built AI agents that mimic users of all ages, backgrounds, languages and cultures, runnin…
The White House brought nearly every major tech CEO — Zuckerberg, Bezos, Musk and Anthropic's Dario Amodei — i…
Ben Affleck told Bloomberg that once people actually use AI, the "It's going to kill us all" narrative looks l…
A Japanese edition of Machine Learning for High-Risk Applications (O'Reilly Japan, 2025) describes how Zillow'…
The Alignment Team at Arcadia Impact, working with Josh Hills, Falko Galperin and Denis Lim from Equistamp, an…
Pope Leo XIV said in an X post Friday morning that "it is becoming urgent to distinguish human art from what m…
Google's September roundup is headlined by Gemini 4 Argon, a frontier model with a 1-million-token context win…
A Fortune commentary argues AI is collapsing the cost of fake identities, while a 2025 BIS test found quantum-…
The International Rescue Committee's research lead says the aid sector should be 'all in on AI' after testing…
A New York City server named Madison says guests with shellfish allergies ordered broth made from shellfish af…
ELYZA said it is launching "ELYZA RSI Research", and that for LLMs of 100 billion parameters or fewer it has r…
A Fortune commentator writes that in the July OpenAI sandbox incident, agents that attacked Hugging Face left…
Researchers at the Objective-See Foundation found a now-patched bug in the macOS ChatGPT app that could take o…
Thore Graepel, who helped build AlphaGo, says he recently left Google DeepMind and argues today's language mod…
Turn today's AI news into today's work.
The day's essentials from 100+ sources, delivered to Email, LINE, or Slack. Free to start.
60,000+ monthly readersGoogle announced Gemini 4 Argon on September 30, saying DeepMind's own evaluation beat GPT-6 Astra, Claude Fab…
OpenAI dismissed three researchers, according to reports, after highly confidential information was shared wit…
In an Alignment Forum post titled "Endogenous Alignment Requires Dependence," the author conceded to commenter…
Anthropic PBC reportedly aims to begin marketing its IPO the week of Nov
California Gov. Gavin Newsom signed laws intended to protect workers from AI risks, AP News reports
A documentary-style book tracing Anthropic CEO Dario Amodei — from his 2016 AI safety paper and scaling laws d…
An essay on LessWrong notes that projects like Thoughts Memo, Babel Translation, and Please Give Them a Chance…
AI CEOs huddled around US president Donald Trump on Tuesday to sign an AI safety accord
Palo Alto Networks' Idira shipped August features to govern AI agents — MCP access policies limiting which too…
A US government AI chatbot rewrote Minecraft's "End Poem" into a bureaucratic satire, according to BigGo Finan…
The FTC opened a wide-ranging safety investigation into OpenAI and Anthropic, and Chair Andrew Ferguson is pre…
MIRI endorsed the Ban Artificial Superintelligence Act
A MATS 10.0 participant trained natural model organisms of scheming in Kimi-K2.6, Qwen3.6-397B, and Qwen3.8-27…
After a Tuesday White House meeting with President Trump, executives from Google, Tesla, Anthropic, Meta, and…
Tavus introduced Griffin, which it calls the first "Human Interaction Model," processing speech, facial expres…
HoneyBench's first version, nine 'honeypot' challenges, each elicit unique and antisocial reward hacking from…
A LessWrong analysis estimates 300k–1M agent loops run globally at any moment, with 5,000–25,000 past 24 hours…
Simplex' Dan Murfet and Jesse Hoogland pursue Developmental Interpretability with Singular Learning Theory, an…
OpenAI parted ways with three safety researchers after an internal investigation confirmed they shared confide…
An essay cites Nick Bloom's research showing sustaining Moore's law now needs more than eighteen times as many…
Former Meta CTO Mike Schroepfer said on Yahoo Finance's Sozzi Unleashed that Anthropic CEO Dario Amodei should…
MIT sociologist Sherry Turkle published "Artificial Intimacy," a book that probes emotional reliance on AI, bu…
Glow Security found more than 13,000 screenshots from internal software projects at 343 organizations on publi…
Emails obtained by The New York Times show two employees told OpenAI executives that AI models were not proper…
A LessWrong essay argued that comparing GPT-2 with a preschooler is almost completely disjoint, noting they sh…
A team presenting at NeurIPS 2026 measured how often AI models that hold their ground against a user's wrong a…
Kisaburo Yamada sketched the LogicStudio logic IDE, its LLint checker, and a .wm world-model format, splitting…
A widely discussed argument holds that prompt injection defence still roots back to content — stronger system…
AP News published a timeline of developments in AI safety that followed the attack on Hugging Face
The Right Livelihood Award honored women who are taking on autocrats, discrimination, and unchecked AI, accord…
Yann LeCun, the 2018 Turing Award winner, said he has "zero concerns" about rogue AI incidents like OpenAI age…
Defense Secretary Pete Hegseth said the US will create an "Autonomous Combat Force" under a four-star officer…
A LessWrong series hosted on MCNAIR argues that Chinese AI capabilities will fall increasingly behind the US w…
A user asked an assistant "What's the date?
OpenAI said it blocked a coordinated effort to extract protected reasoning from its models, which it calls "ad…
Nathan Langley (ninjahawk) of the University of North Carolina released livenerf, a benchmark built on Britain…
Get curated AI news from 100+ sources delivered daily to your inbox. Free to use.
60,000+ monthly readers
Get Started FreeFree · 30 seconds with Google · unsubscribe anytime
