Inside the frantic 24 hours: Anthropic CEO Dario Amodei rebuffed officials' requests to pull Fable 5 voluntarily before the White House imposed export controls - Politico

Geek, interrupted: The White House imposed export controls on Anthropic's Mythos after suspicions that a China-linked group had accessed the model. - Semafor

absolute muppets

Le wake-up call: European politicians call Anthropic's global shutdown a "wake-up call" and urge investment in homegrown AI alternatives like Mistral - euronews

Two days after Dario Amodei called for mandatory frontier AI testing, the US ordered Anthropic to cut off foreign access to Fable 5 and Mythos 5. I consider how Washington's improvised response exposed the lack of an AI safety playbook. - Tech Policy Press

On balance, does not seem a decision made by competent people on on the merits. People will forgive you for the biggest screwups, but will never forgive you for pointing out their own screwups.

An issue with the death penalty is, if you say all child molesters get executed, makes a kid less likely to report a relative, more likely to recant.

Somewhat related, when the Administration is amateur hour, why even share findings about issues, if all it does is fuel culture war vendettas.

Anthropic co-founder warns AI industry has 'no brake pedal' as Trump's 30-day review window draws skepticism - Yahoo

seems like they found a brake pedal

SpaceX debuted at a $2.1 trillion market cap, the largest IPO in history, signaling investor appetite for Anthropic and OpenAI's planned 2026 listings - Yahoo

SpaceX opened at $135 and closed at $160.95 on debut; Nasdaq changed rules to add it to the Nasdaq 100 in 15 days - Ars Technica

CoreWeave, Nebius, Astera Labs, Rocket Lab, and Teradyne join the Nasdaq 100 on June 22; Nebius has surged nearly 200% year-to-date on AI cloud demand - Yahoo

Pie in the very high sky: SpaceX's $1.75 trillion IPO valuation rests on two unproven bets: a million orbital AI satellites and a reusable Starship - Scientific American

Passive investing turns aggressive: Index fund investors will soon own SpaceX shares involuntarily; economist Burton Malkiel says stick with indexing anyway - The New York Times

Beef Wellington: Wellington Management's Brian Barbetta says AI stocks still underestimate the technology's long-term potential. - Barron's

Hawley and DeSantis lead a GOP revolt against Trump's AI deregulation push, with Republican primaries seeing candidates back data center moratoriums - Politico

Opposition to data centers is surging across 45 states — and Pew finds that the more Americans learn about them, the more they oppose them - Business Insider

UK orders Apple and Google to detect child nude images and launches AI defence taskforce alongside £1.1bn hardware investment at London Tech Week - the Guardian

Don’t cry for me: Argentina considers letting AI-run companies with no human staff gain legal status, own assets, and donate to political campaigns - The Telegraph

KPMG withdrew its report on agentic AI after GPTZero found only 5 accurate citations out of 45, with companies denying the report's claims - Engadget

How Texas trial lawyer Mark Lanier used AI platform Boodlebox to prepare for and win a $6 million verdict against Meta and Google. - Business Insider

Hacking mad: Meta employees revolted against Zuckerberg's July hackathon — unpaid effort that won't count toward performance reviews on top of post-layoff workload - WIRED

Bent bobby busted for bot bunkum: Derbyshire police officer is under investigation for allegedly using AI to fabricate evidence in multiple criminal cases. - Sky News

Forty economists and policy experts met at the Peterson Institute to game out AI's impact on U.S. jobs and society by 2030. - The Wall Street Journal

AI was the top cited reason for U.S. job cuts in May for the third straight month, with 88,000 AI-attributed cuts through May 2026 - Sherwood News

Messy middle management: Labor researcher Molly Kinder says AI's "messy middle" will hit highly paid knowledge workers hardest and calls for targeted policies, not universal basic income - Platformer

Half of UK workers fear AI will take their job as agent adoption soars 90%, though WEF projects net job gains by 2030 - TechRadar

Stanford lecturer argues every company should hire an "AI workflows" specialist to automate internal processes, a role paying up to $183,000 - Yahoo

Forward deployed engineers — AI integrators embedded in client systems — are now tech's hottest hire, with Google posting 59 openings at once - Tech Times

AI adoption fails when companies apply new technology to old org structures, a new book argues - Fast Company

Tech worker Jae Park left her sales enablement job at 32 to pursue carpentry after AI reshaped her workflow. - Business Insider

A randomized study of 1,222 participants finds brief AI use measurably reduces cognitive persistence and unassisted performance. - Psychology Today

Software prices rose 14.5% year-over-year in May, the largest annual increase on record since 2000. - Benzinga

A $200/month AI subscription can buy $8,000 in API costs, driving firms to DeepSeek and open-source alternatives - Tom's Hardware

Gemini Spark, Google's $99/month AI agent, leads on data access but fails on reliability and largely duplicates free Gemini features - PCMAG

Firefox added an AI kill switch after user backlash, but only 1% have used it; CEO warns AI-first browsers could close off the web - CNET

AI 'nudify' tools are fueling a surge of deepfake sexual images targeting minors at school - The Wall Street Journal

UK MP Jess Asato filed a High Court claim against xAI after Grok allegedly generated nonconsensual deepfake images of her. - Metro

Adaptive Security raises $140M and enlists Conan O'Brien for enterprise training videos on AI-driven phishing and deepfakes - TechRadar

Russian families are paying AI creators to generate videos and photos of soldiers killed or missing in Ukraine, provoking ethical debate. - BBC

Open tab for open source: OpenAI launches Codex for Open Source, offering maintainers six months of ChatGPT Pro and API credits for coding and maintenance tasks - OpenAI

Zhipu releases GLM-5.2, an open-source model with a 1M-token context window, framing it as a response to frontier model access restrictions - X (formerly Twitter)

Stanford researchers position TextGrad as an extension of DSPy and ProTeGi, generalizing textual gradients from prompt optimization to instance-level optimization across diverse domains. - Hacker Noon

Thirty mathematicians at Harvard evaluated AI math skills, raising questions about whether machines can replace researchers in the field - The Washington Post

Hollywood directors secure AI protections and a 24.4% health plan boost in new 4-year studio contract - Deadline

Gore Verbinski proposes film AI ratings at Taormina, arguing scripts written by AI should earn an 'F' - Variety

DIY or Die: Ukrainian UAV expert says off-the-shelf drones cannot survive electronic warfare and urges every military unit to build its own drone lab. - TechRadar

Follow the latest AI headlines via druce.ai on Bluesky