State capture as a service: US officials have discussed taking government equity stakes in major AI companies after OpenAI CEO Sam Altman pitched the idea. - The Wall Street Journal

unlimited cash plz

The dispute between the Trump administration and Anthropic may be thawing as the company prepares for a $1 trillion IPO. - Reuters

Anthropic says slowing AI development sprints would benefit the world, the same week it filed an IPO prospectus ahead of OpenAI. - theregister

Claude now authors more than 80% of Anthropic’s production code, with engineers shipping 8x more per day as AI task duration doubles every four months - Anthropic

Anthropic deployed an automated Claude code reviewer in its CI/CD pipeline that caught roughly one-third of bugs behind historical outages on claude.ai - VentureBeat

Anthropic, OpenAI, and SpaceX are racing to go public, with combined AI IPO valuations potentially topping $3 trillion - WIRED

Hedging your AI bets: About 90 investment firms hold stakes in both OpenAI and Anthropic, with 42% of OpenAI investors also backing its rival. - WIRED

Anonymous MAGA memo alleged there was a backdoor in AI chip export restrictions, causing a ruckus and closing of loopholes that may or may not have existed. - Bloomberg

Maple eh-AI: Canada is pursuing domestically developed sovereign AI to reduce dependence on US technology platforms. - theregister

The EU proposes barring non-European cloud providers from sensitive public sector workloads and creating a “Eurocloud” data-sharing forum across 27 member states - 4sysops

Pentagon runs AI-generated propaganda site La Tilde targeting Latin American audiences on behalf of US Special Operations Command. - The Intercept

NOAA’s AI cuts weather warning turnaround from one hour to 5–7 minutes; the National Hurricane Center now issues AI-translated Spanish advisories for Atlantic hurricanes - Watts Up With That?

Rosalind Frankenstein: OpenAI releases biodefense action plan alongside GPT-Rosalind and Rosalind Biodefense to strengthen biological security, after AI CEOs co-sign letter urging tech safeguards. - OpenAI

OpenAI rolls out upgraded ChatGPT memory using a new dreaming architecture with a user-facing summary page for review and control. - OpenAI

OpenAI launched Sites, a Codex feature that lets business users build and share internal apps from plain-text prompts. - Inc.

Apple is still tagging the new Siri as “beta” internally, and may gate September access behind a waitlist. After all this time, Siri-ously? - MacRumors

WWDC leaks point to an M5 Mac Mini with local AI acceleration, Apple TV 4K with A17 Pro, and a 7-inch Smart Home Hub. - Geeky Gadgets

Microsoft’s AI chief aims to eliminate Anthropic spending and claims MAI-Thinking-1 matches Claude on coding benchmarks at lower cost - TNW

CheapSeek: US tech firms are rapidly adopting DeepSeek, which topped Ramp’s May breakout SaaS rankings, as its lower costs undercut OpenAI and Anthropic - TechRadar

DeepSeek’s terms require user data to be stored in China, where companies must cooperate with state intelligence requests - 9to5Mac

Vax machine: Cambridge researchers trialed the first AI-designed vaccine antigen in 39 humans, targeting all coronaviruses including potential pandemic strains. - BBC

oooh AI vaccines will surely not trigger anyone

Russia's personalized cancer vaccine program has enrolled over 40 patients, using AI to identify tumor mutations and produce custom peptide vaccines in 42 days. - RT International

FDA accepts AI-based digital liver model for review under its drug development tool qualification program. - Yahoo

can it model a nice chianti to go with that

SoftBank’s Masayoshi Son calls AI “50x bigger” than the dot-com boom, pledging $87 billion for France AI infrastructure, while JPMorgan and BofA warn of bubble conditions. - Yahoo

Bloomberg MLIV investors debate whether the AI trade has run too far, with the boom's end seen as inevitable but timing uncertain. - Bloomberg

Amazon engineers call out $200B data center spending after 30,000 layoffs; Seattle passed a one-year construction moratorium and 14 states are weighing bans - Yahoo

Token burn notice: Hudson River Trading AI head Iain Dunning describes the firm's heavy AI token spending, compute bottlenecks, and plans to develop proprietary chips. - Bloomberg

Cast AI data shows 95% of enterprise GPU capacity sits idle, while AI startups remain priced out of compute access. - TechRadar

Supabase raised $500 million at a $10.5 billion valuation, with AI coding tools — Claude Code chief among them — now driving most new database deployments - CNBC

Alibaba open-sources its internal AI code review CLI, which has served tens of thousands of developers and flagged millions of defects. - GitHub

Sync or swim: Backup failures in the rsync open-source project were traced to AI-assisted commits, devs lose their cool in the debate over vibe-coding in critical infrastructure - theregister

Developer advocates inserting a policy-aware API gateway between enterprise AI agents and internal systems to enforce scoped access and audit trails. - Hacker Noon

Tutorial explains how to build a warm handoff from an AI voice agent to a human using AssemblyAI and Twilio, preserving full call context. - Hacker Noon

Google accidentally exposed a Troubleshooting mode in Gemini that uses interactive widgets and lower temperature to hand-hold users through problems. - Android Authority

Our Ai code sucks: Google employees share internal memes mocking AI code quality even as CEO Sundar Pichai touts AI-generated code. - 404 Media

A hidden Google Chrome Canary flag routes search queries directly to AI Mode instead of traditional results pages; Google calls the feature exploratory. - Engadget

Wired investigation finds Meta quietly installed facial-recognition “NameTag” code on tens of millions of devices; Meta calls it “exploration” - Mashable

Meta's smart glasses companion app Stella contains a dormant but functional end-to-end facial recognition pipeline, a researcher found. - Buchodi's Threat Intel

AI-powered facial recognition now exceeds 99% accuracy in controlled settings and is expanding across airports, arenas, and public transit. - Fast Company

New claimants join UK Labour MP Jess Asato's lawsuit against Elon Musk's xAI over non-consensual sexual images generated by Grok. - the Guardian

Lemon AIeed: AI image tools now produce undetectable fraudulent car listings with 20 photos in 3 minutes; Deloitte estimates AI-enabled fraud hits $40 billion by 2027 - Forbes

New AI system detects fake online reviews with 94.2% accuracy by fusing text analysis, reviewer behavior signals, and image inspection. - Digital Trends

YouTube, Instagram, and TikTok label AI content but offer no filter toggle, and none of the major platforms have plans to add one - The Verge

Peptide and HRT vendors are seeding Reddit with “answer engine optimization” posts engineered to surface in ChatGPT and Google AI answers - TechSpot

53% of consumers distrust AI-generated search results, organic traffic is down 5–35% for some brands, and AI detection tools disproportionately flag non-native English speakers - Fortune

Training bots now account for 50.6% of web traffic vs. 10.7% for search bots, generating no referral traffic for publishers while raising hosting costs - PPC Land

AI bots now account for 57.5% of internet traffic — ahead of Cloudflare CEO’s 2027 prediction — as agentic crawlers overwhelm small website hosts - Tom's Hardwar

NPR/Ipsos poll finds nearly 3-in-4 K-12 teachers believe AI will have greater impact on education than the internet. - NPR

AFT president Randi Weingarten calls for banning AI in elementary schools and limiting screens through second grade to address children's attention crisis. - The 74 Million

It’s No Wonder Grads Are Booing Their Commencement Speakers - The New York Times

AI has already changed writing. Now the technology is changing what it means to read. - The Atlantic

Psychologist Gloria Mark warns that delegating writing and evaluation to AI chatbots erodes critical thinking, emotional intelligence, and sense of purpose over time - MIT Technology Review

Gen Z hostility toward AI is rising sharply, with excitement dropping from 36% to 22% and anger climbing to 31% in one year. - Fast Company

The word 'monoculture' is being used simultaneously to mourn the loss of shared cultural experience and to criticise AI-driven cultural homogenisation. - Fortune

Brookings analysis finds AI-exposed workers are concentrated in Democratic-leaning counties, making AI backlash both an asset and risk for Democrats in November. - Fortune

Think-tank CEO Daniel Schreiber proposes funding a universal high income of $60,000 per year to offset AI job losses. - Bloomberg

Walmart shareholders voted against a proposal requiring the company to report on how AI affects workforce well-being. - Reuters

A software engineer won a religious accommodation to avoid using AI at work; Pope Leo XIV’s encyclical may inspire more faith-based exemptions - Business Insider

Philosophers argue undisclosed AI use in communication is deceptive and morally wrong in non-trivial contexts. - The Conversation

U.S. farmers are adopting AI tools including robotic milking systems, laser weed zappers, and livestock activity monitors to offset labor shortages. - The New York Times

Experts warn that viral humanoid robot demos are cherry-picked and mask significant gaps in real-world generalization and reliability. - Ars Technica

Grand theft auto-nomous: thief used a Waymo robotaxi to steal yoga clothes from a San Francisco studio in January and remains at large. - Engadget

Goodfellas, bad look: Director Boots Riley criticized Martin Scorsese's advisor role at AI firm Black Forest Labs, suggesting Scorsese took it mainly for the money. - Deadline

Founders Fund launched a Mafia game show featuring Sam Altman, Palmer Luckey, and other Silicon Valley figures. - Business Insider

ESPN is reviewing its use of AI-generated promotional images after a moving portrait of Tony Parker drew widespread backlash. - Front Office Sports

why not AI-generate the whole game and take bets

Trump, Abbott, and Blakeman posted AI-generated basketball images depicting themselves dunking on New York Governor Hochul. - Gothamist

Papal bull: fact-checkers find unproven but statistically suggestive evidence that Vatican staff used AI to help draft Pope Leo XIV's anti-AI encyclical. - Snopes

Writer uses ChatGPT for home DIY projects and finds AI gives overconfident, sometimes dangerously wrong advice. - the Guardian

Follow the latest AI headlines via druce.ai on Bluesky