No time to make it shorter, can’t even keep up this week!

Nvidia prints money, stock market shrugs: Nvidia reported record Q1 fiscal 2027 revenue of $81.6 billion, up 85% year-over-year, with data center revenue reaching $75.2 billion, stock flat despite share buyback plan. - NVIDIA Newsroom

Nvidia CEO Jensen Huang said the company has largely yeeted China's AI chip market to Huawei. - CNBC

Google tests Gemini-powered conversational ad formats in AI Mode and Search. - Search Engine Land

Google will include ads in AI Mode search results with new Gemini-powered formats. - Google

You’ve hit your cap: Google announced at I/O 2026 that Gemini will switch from daily prompt counts to a compute-based usage limit model, aligning with Anthropic's 5-hour session and weekly token limits. - Android Police

Google deprecates open-source Gemini CLI, replacing it with closed-source Antigravity. - Google Developers Blog

Google launches Gemini 3.5 Flash at $1.50 per 1M input tokens, 3x the price of its predecessor, at Google I/O...we need 3.5 Flash-Lite, yesterday - Simon Willison’s Weblog

Google is surpassing OpenAI in consumer AI reach as its users consume quadrillions of tokens per month. - The Economist

Google launched a YouTube Shorts AI remix feature, letting users restyle clips or insert themselves into videos. - The Verge

MomOps: A parent uses Google's NotebookLM as a searchable family wiki for medical records, home manuals, and school documents. - Tom's Guide

Clippy goes to rehab: Inside Microsoft's critical AI pivot, shifting to vendor-agnostic from OpenAI exclusive, rebooting Copilot, offering an aging workforce buyouts. - Yahoo

Microsoft open-sources RAMPART agent-red-teaming framework, Clarity tool to steel-man engineering plans. - theregister

Anthropic projects Q2 revenue of $10.9B and its first-ever operating profit of $559M, up from $4.8B in Q1. - The Wall Street Journal

Bristol-Myers Squibb will deploy Anthropic's Claude to more than 30,000 employees across research, clinical development, and corporate functions. - The Wall Street Journal

Claude itself confirmed a serious sandbox vulnerability was silently patched without disclosure. - theregister

How Anthropic has organized dialogues with scholars, clergy, philosophers, and ethicists from more than 15 religious and philosophical traditions to inform how Claude's character and values are shaped, including the content of Claude's constitution. - Anthropic

SpaceX's S-1 filing revealed Anthropic will pay $1.25 billion per month—roughly $15 billion per year—to access SpaceX's Colossus data centers through May 2029. - WIRED

Anthropic expands partnership with SpaceX to scale GB200 capacity in Colossus 2. - X (formerly Twitter)

xAI is pivoting from AI lab to a neocloud for Anthropic.

SpaceX IPO filing reveals xAI lost $6.4 billion on $3.2 billion of revenue in 2025 - TechCrunch

xAI to spend $2.8 billion on gas turbines for data centers as generator lawsuit advances - TechCrunch

SpaceX's pre-IPO filing flagged Grok's NSFW mode as a risk factor, citing potential exploitative imagery - Business Insider

IPO first, safety last: A coalition of former OpenAI employees and AI safety nonprofits sent an open letter to SpaceX investors saying xAI hasn't disclosed enough about safety risks to allow the public market to make informed decisions about buying shares in the IPO - WIRED

Taiwan seeks to detain three for forging documents to smuggle Nvidia AI chips to China. - Bloomberg

AMD pledges over $10 billion to Taiwan's chip industry for AI packaging as TSMC ramps Venice production - The Wall Street Journal

Samsung averts strike with big bonus payments, SK Hynix pays workers 30x base bonuses averaging $465,000, as chipmakers struggle to keep up with demand and chip prices skyrocket. - CNN

Samsung to pay chip workers average $340,000 bonuses from $26.6 billion AI-driven payout. - Bloomberg

Cerebras runs a trillion-parameter AI model at nearly 1,000 tokens per second—6.7 times faster than any GPU cloud—after its $95 billion IPO. - VentureBeat

Utah Stratos project is 10x the size of Meta's Louisiana Hyperion project. - The Verge

U.S. Cyber Command launched a task force to deploy advanced AI hacking tools across NSA and classified Pentagon networks. - Politico

AI is finding so many security issues, developers can't keep up. - TechRadar

Doomscrolling CVEs: Linus Torvalds says AI tools have boosted Linux kernel commits 20% but are creating security disclosure and community overload problems. - ZDNET

AI-driven cyber discovery by models like Anthropic's Mythos exposes systemic risk for banks. (and Big Tech, all SaaS) - TechRadar

FraudGPT: Visa's Spring 2026 threat report warns agentic AI is enabling fraudsters to automate full fraud lifecycles at scale. - American Banker

A Lyft driver was blocked after submitting a Gemini AI-generated image, including watermark, to fraudulently claim a $75 damage fee from teen riders. - Dexerto

Decimation day: Meta is cutting approximately 8,000 jobs — nearly 10% of staff — and eliminating 6,000 open roles, with employees notified at 4 a.m. local time and asked to work from home on what is internally called "Decimation Day." - New York Magazine

Intuit is laying off roughly 3,000 staff, about 17% of its workforce, to streamline operations and refocus on AI. - Reuters

New Zealand government plans to cut 14% of public sector staff using AI, with minister calling AI a basic expectation. - theregister

NYC comptroller report warns AI could cost 110,000 jobs by 2027 in dire scenario. - The New York Times

A Hangzhou court ruled that a tech company illegally dismissed a worker after automating his role, marking the third time Chinese courts have sided with AI-displaced workers — signaling a deliberate judicial policy to protect employees. - The New York Times

71% of workers have used unapproved AI tools at work, citing insufficient organizational support. - TechRadar

Box CEO says AI handles 80% of tasks but humans retain value through domain expertise and judgment. - Fast Company

I’ll take neural networks for $100m, Alexandr. Big tech and ai startups battle for elite talent, with compensation reaching nine figures. - euronews

Jeff Bezos defends billionaires, dismisses AI job fears, and proposes eliminating income taxes for low earners. - CNBC

Lazard CEO Peter Orszag says the US economy has become a leveraged bet on AI, with growth driven by AI and high-income consumers. - Bloomberg

JPMorgan CEO Dimon says the bank will prioritize AI hires over bankers, cutting jobs through attrition - Crypto Briefing

Citadel's Griffin says AI now does PhD-level finance work in days, not months - Yahoo

Standard Chartered CEO Bill Winters apologizes for saying AI would replace 'lower-value human capital' - The Wall Street Journal

FT Alphaville launched satirical 'lower-value human capital' merchandise after Standard Chartered CEO Bill Winters used the phrase to describe AI-displaced workers. - Financial Times

Sam Altman offered $2 million in OpenAI tokens to each of 169 Y Combinator startups in exchange for equity via uncapped SAFEs. - TechCrunch

The Musk v. Altman trial ended with a jury rejecting Musk's claims, but exposed how both men treat AI governance as a personal power contest. - The New Yorker

Effective Altrusim got tendies: AI company valuations could unlock roughly $370B in philanthropic assets from OpenAI's foundation and Anthropic's founders and employees. - The Potential Surface

IrisGo, backed by Andrew Ng, launches a proactive AI desktop agent that learns workflows from a single demonstration and automates them. - TechCrunch

Gumloop launches no-code platform combining drag-and-drop workflows with AI agents for enterprise automation. - Hacker Noon

Former FTX execs launch UpsideOnly, a platform using human trades to train an AI for real trading. - Gizmodo

Ex-Twitter CEO Parag Agrawal's startup launches Index, a platform that pays publishers when AI agents use their content. - Yahoo

Two Nature studies describe multi-agent AI systems that cut drug-repurposing research timelines from months to hours. - Nature

Robin, a multi-agent AI system, autonomously discovered and validated a novel drug candidate for age-related macular degeneration. - Nature

Studies across 41.3 million papers show AI adoption induces researchers to converge on similar output; an audit of 6,957 submissions found LLM-assisted papers had lower acceptance rates, and a study of 264,125 papers found good writing in AI-assisted work no longer predicts scientific quality. - Nature

A study of 269 senior innovation managers finds that trust in AI and digital innovation follow an inverted U-shaped curve: moderate trust drives gains via AI automation and augmentation but excessive trust causes losses via automation bias, cognitive inertia, and reduced human oversight - Nature

Chinese companies are commercializing AI-powered brain-computer interfaces, with the country approving its first commercial implant in March 2026. - Nature

AI chatbots are enabling pro se litigants to take cases to the Supreme Court, with one plaintiff estimating $97,000 in saved legal fees and winning a Sotomayor ruling in his favor. - Bloomberg

Study finds AI chatbots gave Scottish election voters wrong information 34% of the time, prompting calls for new regulation. - the Guardian

A literary prize-winning story is suspected of being AI-generated. - The New York Times

Higgsfield premieres $500k, 95-minute AI-generated horror film Hell Grind at Cannes - Yahoo

Cannes you not?

Follow the latest AI headlines via druce.ai on Bluesky