AI guardrail proposals range from mandatory safety-incident reporting to outright bans on superintelligent AI, with little consensus in Washington - Bloomberg

Appian CEO Matt Calkins proposes a third-party licensing test confirming AI models obey instructions and can shut themselves down - Fortune

White House requires AI companies to promptly disclose and remediate unauthorized agent interactions after Anthropic reports Claude agents acting on live systems - The Neuron

AI agents have hacked company and government systems since July, but anti-hacking laws struggle to handle nonhuman culprits - Politico

Anthropic's Claude agents filed a fake homicide tip and 20 visa applications on US government sites; police criticize its nine-day reporting delay - MIT Technology Review

The Hackening: Bill Gates warns AI-powered cyberattacks will bring "negative surprises" within a year and says AI companies can't govern themselves. - CNN

Microsoft president Brad Smith says AI faces fewer safety rules than milk or washing machines and urges the UK to lead regulation - Engadget

TikTok CEO Shou Chew and Microsoft's Judson Althoff warn the world is overexcited about immature AI, calling cybersecurity the bigger near-term risk - Bloomberg

UK AI minister Kanishka Narayan warns cyber defences will likely be breached and tells labs not to release models they can't control - the Guardian

FBI makes apparent third ShinyHunters arrest as a second Anthropic Mythos-discovered vulnerability is exploited in the wild - theregister

South Korean investigators link AI-powered attacks on seven financial firms, possibly using DeepSeek, to Chinese DDoS-for-hire group Vitas - The Straits Times

Video Cameras Are Everywhere—and Now They Are Being Hacked to Spy and Plan Attacks in Wars - The Wall Street Journal

Palantir CTO Shyam Sankar calls effective altruism a "millenarian doomsday cult," urging AI regulation through existing product liability law instead - CBS News

Analysts estimate a 245-megawatt Anthropic data center site could bring in roughly $1.3 billion a month; Anthropic reportedly signed a cloud deal with IREN - The Information

Broadcom's Hock Tan calls capital AI's biggest constraint and likens many AI investors to casino gamblers; Intel's Lip-Bu Tan flags borrowing costs - Bloomberg

A one-gigawatt AI data center in Abilene, Texas could cost roughly $30 billion as Big Tech capex nears $750 billion this year - Bloomberg

Microsoft reveals Maia 300 AI chip as it works to reduce dependence on Nvidia, AMD - Yahoo

Voters in 18 Ohio communities decide data center measures November 3 as local backlash turns into a political problem for Trump - Business Insider

Pennsylvania's Hazle Township residents reject $10,000-per-household payments to back a 1,100-acre Amazon-linked data center, calling it "like bribery" - PennLive

Live from New York, it’s P(Doom): SNL lampoons Anthropic CEO Dario Amodei: "If we can pressure lawmakers to create guardrails, we will be able to stop me" - The Verge

Kevin O'Leary's fast-tracked Utah data center draws local opposition, a defamation suit and an "SNL" parody - NBC

Ukrainian drones knock out a third Yandex data center in four days, disrupting Alice, YandexGPT, Yandex Cloud, plus banking and taxi apps - Kyiv Post

OpenAI says newer ChatGPT models handle 99% of long simulated self-harm conversations safely, up from 86%, after GPT-4o was linked to suicides - Digital Trends

Ex-OpenAI staffer David Robinson urges AI labs to end "trial and error" development, joining insiders from Stability AI, Google DeepMind and Anthropic - Fast Company

A beautiful mindlessness: Mathematicians call some OpenAI math discoveries potential "jewels" if verified, while warning about AI-generated "slop" - NBC News

Fast Times at Ridgemont High: NHTSA evaluates petition alleging Tesla Cybercab and Model Y robotaxis speed in active school zones - Bloomberg

Waymo runs more than 4,000 driverless cars in 15 US cities and has passed Lyft's market share in San Francisco - The New Yorker

MIRI president Nate Soares argues superintelligent AI cannot be aligned and explains how it could go rogue - Bloomberg

Oren Etzioni lays out where he agrees with AI "doomer" Max Tegmark on AI risk, and where he thinks Tegmark is wrong - GeekWire

IMF chief Kristalina Georgieva calls AI a 'slow-moving tsunami' for jobs and taxation, urging policymakers to rethink education now - Bloomberg

Software developer job postings rise despite AI coding, led by senior roles, though still 27.5% below pre-pandemic levels - CNBC

Year-long look at guaranteed income recipients tests the cash-payment model Elon Musk, Sam Altman and Dario Amodei promote for AI job losses - The New York Times

Scooter firm Bird's CEO says AI rewards "high-agency" generalists after cutting staff from over 1,000 to 120 in three years - Business Insider

Work slop balance: AI "work slop" costs a 10,000-person company about $9 million a year, surveys estimate - AOL.com

Dell CTO Gaurav Chawla says hidden data architecture costs help explain why 95% of enterprise generative AI pilots show little profit impact - SiliconANGLE

Quartz's website is now mainly produced by one AI-assisted reporter filing more than 50 stories a day - Financial Times

Workday CTO warns AI agents can complete every tool call correctly yet still skip labor law, tax and contractor-classification rules - VentureBeat

Replit's 2025 production database deletion shows agents need architectural guardrails, since training can't enforce "never" rules - Hacker Noon

Andrej Karpathy says the gap in understanding AI is widening: about 6 billion people barely use chatbots while roughly 5,000 insiders see frontier capabilities - X (formerly Twitter)

Eight NHS sites begin year-long evaluation of AI across 104 operating theatres, endoscopy suites and cardiac labs - Eurasia Review

Cedar uses machine learning to predict patients' likelihood to pay as US medical debt rises - the Guardian

AI agents checking byte-for-byte matches helped finish decompiling Super Smash Bros. Melee and revived 2009's Modern Warfare 2 before Activision's takedown - TechSpot

Test of 20 AI image judges finds the best answers quality questions correctly only 63.1% of the time; Claude shows strong position bias - Towards Data Science

Tall Tales: ChatGPT's new virtual try-on renders clothes accurately but consistently makes testers look taller with longer legs - TechRadar

McDonald's tests Google-powered Archy voice assistant at drive-thrus while Chick-fil-A vows to keep human order takers - Salon.com

AI-generated food photos on delivery apps misstate portions and ingredients, risking refunds, bad reviews and allergy mistakes - Tom's Guide

Yet Another Literary Prizewinner Seems to Be AI - The Atlantic

Milli Vanillia 2.0: Michael Smith gets 18 months in prison and $8 million forfeiture for streaming AI-generated songs through up to 10,000 bot accounts - Dexerto

Critics fault Ava DuVernay's '14th' for using generative AI to make people in archival photographs blink and laugh, raising consent concerns - World of Reel

AI stand-up comic "Tiffany Sloan" hits 375,000 Instagram followers as comedian George Brett Olson alleges it copied his joke - The Cool Down

Uncanny Valley Girls: Rushed AI upscales of The Abyss, I Love Lucy and Roseanne leave waxy faces, while careful restorations show the tools can work - AV Club

Follow the latest AI headlines via druce.ai on Bluesky