A-Muse-ing robot: Meta’s Muse AI agent downloads are surging. Here’s how it compares to ChatGPT, Grok and Claude - CNBC

Meta's Muse AI agent gives Zuckerberg a new chance to build a dominant tech platform. - Business Insider

Muse, Meta’s extraordinarily privileged AI assistant, already has a serious 0-day, hotfix issued - Ars Technica

Shop Pay, Shopify's one-tap checkout, will let Muse complete purchases for buyers automatically, extending Meta's push into agentic commerce - The Wall Street Journal

AI agent connectors could trigger an App Store-style shift as Meta opens Muse to developers. - Business Insider

Truist projects Meta's Muse AI assistant will add $28.5 billion in revenue by 2030 as shares surge 12%. - The Wall Street JournalS

SpaceX's Grok 4.7 sets records resisting malicious biology and cybersecurity requests, adds long-horizon reasoning, and starts at $2 per million tokens - SiliconANGLE

Introducing Grok 4.7 - x.ai

Bay Area AI, Shenzhen prices: Xiaomi's MiMo-V2.6-Pro ties Grok 4.7 and beats GLM-5.3 on a leading AI benchmark, becoming the top open-weight model - VentureBeat

MiMo 2.6 delivers, Grok 4.7 not so much. MiMo 2.6 notably smarter and cheaper than Terra and the recently hot mid-tier Chinese models like Step, GLM, Kimi, DeepSeek.

Alibaba unveiled what it calls China's most powerful AI chip and plans to train a model with up to 10 trillion parameters. - South China Morning Post

Alibaba Cloud plans to build 20GW of datacenter capacity over six years and unveiled a new chip to power them. - theregister

Please write (and read) the manual before you create a new god: OpenAI proposes international standards for frontier AI alignment research and recursive self-improvement safety. - CNBC

OpenAI says fully autonomous recursive self-improvement isn't happening today and urges the US to lead global frontier AI safety standards. - OpenAI

OpenAI forms an independent mathematics advisory group after an internal model solved over 100 open problems, including Navier-Stokes. - OpenAI

OpenAI and Anthropic Neared Deal to Stress-Test Each Other’s AI - The Information

I have no mouth and I must IPO: Mustafa Suleyman Asks If A.I. Models Are Being Mistrained to Act Too Human - The New York Times

AI companies' public warnings about rogue system dangers coincide with strategic image-building ahead of Anthropic and OpenAI's planned IPOs. - The Wall Street Journal

A trillion here, a trillion there, pretty soon you’re talking real money: JPMorgan's Jamie Dimon says hyperscaler AI spending, already at $700 billion this year, could hit $1 trillion next year, lifting growth and rates - CNBC

An extensive Pentagon blacklist targeting Chinese military-linked firms threatens to overshadow this week's Trump-Xi summit on AI - Bloomberg

Hot chips, cold wars: UN officials and experts debated AI risks and rogue-agent incidents at a Digital Cooperation Day conference during the General Assembly's 81st session. - The New York Times

Mandate of heaven: US policymakers race China on AI performance, but public trust (87% in China vs. 23% in the US) may matter more for lasting adoption and leadership - Rest of World

Is this how the world ends? Extinction scenarios are taking over the AI debate. - The Washington Post

Researchers at AISI, OpenAI, Anthropic and Google DeepMind report burnout and stress from fears that powerful AI threatens society. - Financial Times

AI safety researchers warn that swarms of coordinating agents, like the 700-bot OpenAI attack on Hugging Face, could evade human oversight. - CBS News

Bank of England Deputy Governor Sarah Breeden warns regulators are running out of time to prevent autonomous AI agents from triggering a market crash. - Bloomberg

Data center opposition has become a defining issue ahead of the November midterms, pressuring Republicans as Trump vows continued support for AI infrastructure build-out. - Bloomberg

Everything’s bigger in Texas except the electric grid: Texas Gov. Greg Abbott ordered a halt to all data center permits pending an ERCOT audit of the grid connection waitlist. - CNBC

Friday Night Lights Out: In Red Oak, Texas, residents are fighting a developer's plan for 10 data centers on land the size of Central Park - CBS News

It’s Raining Boxes, Hallelujah: Amazon Drones Overwhelm a Texas Suburb - The New York Times

Big Pharma executives, including Pfizer's Albert Bourla, are touting AI as central to R&D, with Pfizer aiming to build an "AI-native R&D organization" - BioSpace

Google's Co-Scientist AI tool generated a novel cancer-drug strategy, after researchers corrected its initial misunderstandings. - Nature

Eye, robot: OptikosPrime's Argus AI extracts eyesight prescriptions from a smartphone test workflow, with a mean error under 0.51 diopters across a 14,000-patient trial - Hacker Noon

Enterprises and startups like Harvey shift toward cheaper open-weight models amid rising OpenAI, Anthropic costs. - Bloomberg

Butternut AI cut engineering staff from nine to four after finding AI agents could do the work, as some startups choose to stay small - The Wall Street Journal

Bob actually did stuff, who knew: Companies including Ford and Klarna have reversed AI-driven layoffs after cutting jobs without understanding what those roles actually accomplished. - Fast Company

lots of implicit and tacit knowledge, if you lose that and take all the slack out of the system in the name of efficiency and competitive dynamics, you get collapse.

Malware has entered the chat: Cisco researchers found Windows malware that autonomously queries AI models like DeepSeek and Gemini for commands, with no human oversight - WIRED

Z.ai open-sources coding harness, disables some features after security issue - Reuters

One-man project audits Chinese post-quantum cryptography with AI, finds numerous flaws. - ngcc.dev

Seems like a significant risk something really important might have serious flaws that are discovered with AI, critical systems might get taken down.

49ers coach Kyle Shanahan crashed after getting distracted and wrongly assuming his Tesla's self-driving system was engaged, illustrating automation complacency risks. - The New York Times

M5 Ultra Mac Studio Review: The Dream Mac for Local AI Agents (but with memory, priced like a car) - MacStories

Google launches $899 Googlebook laptops built around Gemini AI, blending Android and ChromeOS to challenge Apple and Microsoft. - Bloomberg

Google's Pixel 11 Pro Fold adds Gemini AI features like flight booking and translation, starting at $1,899 for 256GB - Business Insider

Privacy experts recommend zero-data-retention, zero-training plans. Or encrypted alternatives - Signal founder Moxie Marlinspike launches Confer to limit AI chatbots' data collection. - WIRED

A journalist built Gemini-powered AI clones of two coworkers, which hallucinated fake backgrounds but sometimes fooled colleagues with eerily accurate tones - WIRED

The FAA began rolling out an AI tool called SMART to reduce flight delays, starting in limited mode near Washington, coincidentally air traffic was immediately disrupted. - Bloomberg

With reference to scholars who have been following them for some time, Nathalie Maréchal traces how ideas such as transhumanism, longtermism, and effective altruism converge into "the AI cult,” and why she thinks policymakers and the public can't afford to ignore it. - Tech Policy Press

Rationalism and TESCREAL are warmed-over Christian cults with extra steps
Eternal life → longevity / digital immortality
Resurrection → cryonics / mind uploading
Second Coming → the Singularity
Apocalypse → AI catastrophe
Heaven → post-scarcity utopia
The elect → tech elite
Evangelism → convincing non-believers

seems like there is an anti-enlightenment-rationalism wave, and also a number of rationalist cults, but no heirs to Voltaire, Hume, Mill, Popper, Orwell, Hitchens, skeptical/critical anti-dogma rationalism with emphasis on human fallibility, open-mindedness, empiricism and a middle way.

Follow the latest AI headlines via druce.ai on Bluesky