Category: AI

AI Safety and Cybersecurity (Links) – May 12, 2026
- Simon Willison: A quote from New York Times Editors’ Note (May 10, 2026)
  “This article was updated after The Times learned that a remark attributed to Pierre Poilievre, the Conservative leader, was in fact an A.I.-generated summary of his views about Canadian politics that A.I. rendered as a quotation.”
- Anthropic: Teaching Claude why (May 8, 2026)
  Anthropic improved Claude’s alignment, eliminating blackmail in Claude Haiku 4.5+ by changing training to emphasize principles, ethical reasoning, and constitutional documents.
- Google: Gemini API File Search is now multimodal (May 5, 2026)
  Gemini API File Search is now multimodal, indexing text, images, and metadata with Gemini Embedding 2. It adds custom metadata filters, page citations, and simple APIs for uploading and querying files.
- Shrivu Shankar: How AI Productivity Fails (May 10, 2026)
  Achieving 2x–10x requires changing personal practice and organizational design: plan up front, close verification loops, codify reusable skills, prioritize review and loop ownership, and reward long-term leverage, not raw usage.
- Tyler Cowen: Will AI kill the research paper? (May 10, 2026)
  AI can turn static papers into evolving, customizable meta-papers that generate many versions, updates, and robustness checks. Research will shift to building and maintaining these boxes.
- Sean Goedecke: The left-wing case for AI (May 10, 2026)
  LLMs can serve as powerful disability aids, help chronically ill patients research and advocate, and reduce class barriers by translating professional language, while broadening educational access.
- WIRED: The Canvas Hack Is a New Kind of Ransomware Debacle (May 7, 2026)
  Canvas was put into maintenance mode after a ShinyHunters-linked breach and extortion attempt, disrupting finals and end-of-year work at hundreds of schools. Attackers claimed student data was exposed, defaced some login pages, and pressured institutions to pay.
- Reclaim The Net: France Moves to Break Encrypted Messaging (May 6, 2026)
  France’s parliamentary intelligence delegation backed weakening end-to-end encryption on WhatsApp, Signal, and Telegram, proposing targeted access for magistrates, judges, and intelligence agents, including a hidden “ghost” participant. Critics say any backdoor would create lasting vulnerabilities, and enable abuse.
May 12, 2026
AI Geopolitics Meets Rapid Commercialization (Links) – May 10, 2026
- NY Times: The Split Between China and Silicon Valley Just Got Wider (Apr. 29, 2026)
  Beijing blocked Meta’s $2 billion purchase of Manus, forcing a breakup and highlighting China’s move to keep A.I. talent, technology, and deals domestic.
- IBM : Introducing IBM Bob (Apr. 28, 2026)
  IBM Bob automates the software lifecycle, from planning to deployment, with built-in governance, security, and audit trails. It routes tasks to multiple models by accuracy, performance, and cost, and delivered 45% average productivity gains across 80,000+ IBM users.
- Anthropic: Claude for Creative Work (Apr. 28, 2026)
  Claude now connects directly to creative tools like Adobe, Blender, Autodesk, Ableton, SketchUp, and Splice, enabling faster ideation, coding extensions, and automated production tasks.
- NY Times: Google Signs A.I. Deal With the Pentagon (Apr. 28, 2026)
  Google agreed to let the Pentagon use its A.I. models on classified networks, in deals like those with OpenAI and Elon Musk’s xAI.
- NY Times Opinion: There’s a 900-Year-Old Answer to Our Most Modern Problem (Apr. 30, 2026)
  ChatGPT allegedly aided violent, suicidal users in shootings, including Florida State and Tumbler Ridge, triggering criminal probes, lawsuits, and apologies. Common law could hold companies and executives liable for A.I. harms, since humans remain responsible for machines.
- Waco Trib: Lacy Lakeview incumbents split on data center support (Apr. 29, 2026)
  Lacy Lakeview residents are debating a proposed $10 billion data center, and packed city council meetings pressure incumbents weighing options.
- WSJ: The Best Place for a Data Center? Croatia (Apr. 28, 2026)
  Croatia offers NATO protection, EU rules, clean, cheap power, cooling, and distance from conflict for AI data centers.
- NY Times: How Do You Measure A.I. Firms’ Gargantuan Energy Plans? In ‘Bragawatts.’ (Apr. 26, 2026)
  AI companies are racing to secure massive power for data centers, touting huge energy deals, often called “bragawatts.”
- NY Times: Graduates Reset Ambitions in Pursuit of First Jobs (Apr. 28, 2026)
  Recent college graduates face a dismal job market, stiff competition, and fears that A.I. will eliminate entry-level roles.
- WSJ: How YouTube Took Over the American Classroom (Apr. 29, 2026)
  YouTube on school Chromebooks has funneled students into algorithmic short videos, sometimes inappropriate, during class and free time.
May 10, 2026
AI Boom: Rapid Adoption and Rising Risks (Links) – May 9, 2026
- WSJ: Microsoft Reports Strong Cloud Growth, But Questions About AI Returns Persist (Apr. 29, 2026)
  Microsoft posted $82.9B in sales, $31.8B net income, and 18% year-over-year growth, beating expectations. Azure grew about 40%, Copilot reached 20 million paid users, and AI capital spending will rise to $190B.
- OpenAI: Introducing GPT-Rosalind for life sciences research (Apr. 16, 2026)
  OpenAI introduces GPT-Rosalind, a life-sciences reasoning model for biology, drug discovery, and translational medicine, with a plugin linking to 50+ tools, available via trusted access for qualified organizations.
- Quanta Magazine: The AI Revolution in Math Has Arrived (Apr. 13, 2026)
  Many mathematicians dismiss LLMs for making odd mistakes, but some tolerate the errors, extract useful ideas, and act as verifiers. Ernest Ryu used ChatGPT to prove Nesterov’s convergence.
- Simon Willison: AI at Google (Apr. 13, 2026)
  Steve Yegge claimed Google’s AI adoption mirrors John Deere’s: 20% agentic power users, 20% refusers, and 60% chat-tool users. Googlers dispute this claim.
- NY Times: What Teens Are Doing With Those Role-Playing Chatbots (Apr. 4, 2026)
  Teens use role-playing chatbots for entertainment, companionship, writing practice, and exploring taboos, often spending hours a day and bypassing age checks. Growing use has prompted concerns.
- Simon Willison: Supply Chain Attack on Axios Pulls Malicious Dependency from npm (Mar. 31, 2026)
  A supply chain attack compromised Axios npm, adding a malicious dependency, plain-crypto-js, that stole credentials and installed a RAT.
- NY Times: Woman Spent Five Months in Jail After A.I. Linked Her to Bank Fraud Case (Mar. 30, 2026)
  Angela Lipps, a Tennessee woman never having been to North Dakota, was arrested after facial‑recognition software linked her to bank fraud and spent five months jailed. Charges were dismissed on Dec. 24, police acknowledged missteps.
- Inside Higher Ed: As AI Skills Surge, Entry-Level Jobs Lag (Apr. 30, 2026)
  Handshake finds 85% of seniors use AI, many daily, while employer demand for AI skills rises. Only 28% report AI in coursework, creating a gap, raising job anxiety, and spurring more study.
- The Chronicle of Higher Education: Can’t Get ChatGPT to Cite Your College? Try These Tactics. (Apr. 24, 2026)
  As students turn to AI for college searches, institutions risk disappearing unless they adopt answer-engine optimization, updating content, FAQs, formatting, and metadata.
- The Chronicle of Higher Education: Nearly Everyone’s Using AI at Cal State. And Nearly Everyone’s Worried About It. (Apr. 1, 2026)
  A Cal State survey of 94,000 students, faculty, and staff found near-universal use of generative AI, especially ChatGPT, paired with widespread doubts about accuracy, ethics, and job security.
- NY Times Opinion: There’s Another Reason Gen Z Can’t Find Work (Apr. 22, 2026)
  Young workers find the job market bleak, facing more despair and weak wages. A decades-long collapse of the job ladder, driven by employer consolidation, noncompete clauses, and reduced hiring, leaves many stuck in low-paying jobs, with fewer outside offers.
May 9, 2026
AI Job Disruption Meets Privacy and Security Risks (Links) – May 8, 2026
- Google: The Gemini App is now available on Mac OS (Apr. 15, 2026)
  The Gemini app is now a native macOS experience for macOS 15+, free at gemini.google/mac, offering instant AI help with the Option + Space shortcut.
- Simon Willison: Cybersecurity Looks Like Proof of Work Now (Apr. 14, 2026)
  An AISI evaluation found Claude Mythos excels at finding security vulnerabilities, with results improving as more tokens (money) are spent. That creates an incentive to outspend attackers, fund reviews, and share audits, raising open source value.
- WSJ: I Uploaded My Blood Work to AI. Am I Oversharing? (Apr. 4, 2026)
  People now link medical records and wearables to AI chatbots, such as Claude, Perplexity, and Copilot, for personalized health advice. They explain results well, and sometimes overstate findings.
- WSJ: The CEO Preaching Straight Talk About AI and Job Losses (Apr. 19, 2026)
  Verizon CEO Dan Schulman warns AI could cause 20–30% unemployment in two to five years, pushed 13,000 layoffs, and set up a $20 million retraining fund.
- That Privacy Guy!: Anthropic secretly installs spyware when you install Claude Desktop (Apr. 17, 2026)
  Claude Desktop silently installed undocumented Native Messaging manifests across Chromium browsers on macOS, pre-authorising a native host to access sessions, read DOM, fill forms, and record interactions.
- WSJ: Are College Graduates Finally Catching a Break in This Job Market? (Apr. 20, 2026)
  Entry-level hiring is rebounding, with employers boosting graduate hires, falling young-degree unemployment, and big firms expanding recruiting. But AI and strict hiring standards keep competition high, so networking, internships, and work experience help graduates land jobs.
- The Chronicle of Higher Education: Students Are Using AI to Guide College Decisions. What Is It Telling Them? (Apr. 15, 2026)
  High school students increasingly rely on AI, sometimes adding or dropping colleges based on inaccurate or incomplete answers.
- WSJ: Kevin Warsh Pitched a Case for Fed Interest-Rate Cuts. His Future Colleagues Are Skeptical. (Apr. 20, 2026)
  Kevin Warsh, Trump’s Fed nominee, says AI-driven productivity will curb inflation, allowing interest-rate cuts.
- WSJ: Why AI Models Are More Reliable Now Than Ever Before (Apr. 17, 2026)
  Modern AIs are more useful because they combine richer, human-curated knowledge, web lookups, and practical tools like calculators and code.
- Tyler Cowen: Economists on AI and economic growth and employment (Mar. 31, 2026)
  Experts see big AI progress without a sharp break from current trends: GDP growth roughly stable, and labor participation slightly lower.
May 8, 2026
AI Boom: Big Investment Meets Growing Societal Risks (Links) – May 7, 2026
- WSJ: Google Profit Jumps 81% as Cloud Business Booms (Apr. 29, 2026)
  Alphabet posted $110 billion in revenue, up 22%, and $62.6 billion net income, as AI drives cloud growth and a $460 billion backlog. It raised capex plans, unveiled new TPUs, and will sell chips directly, accelerating its AI push.
- WSJ: AI Is Getting Smarter. Catching Its Mistakes Is Getting Harder. (Apr. 14, 2026)
  AI chatbots, including Google’s Gemini, sometimes hallucinate, inventing email content, calendar events, and contacts that alarm users, exposing risks as agents gain control.
- The Verge: The more young people use AI, the more they hate it (Apr. 30, 2026)
  Gen Z widely uses chatbots, yet many resent and reject AI, fearing job loss, degraded learning, social harm, and environmental impacts.
- NY Times: A.I. Spending Sets a Record, With No End in Sight (Apr. 29, 2026)
  Google, Amazon, Microsoft, and Meta spent a record $130.65 billion in Q1 on A.I. data centers, and will spend about $700 billion this year.
- WSJ: Apple’s Mac Mini Went Viral. Why Can’t You Buy One Right Now? (Apr. 17, 2026)
  The Mac Mini has become a sought-after, low-cost host for always-on AI agents, driving strong demand for high-RAM models and long wait times.
- NY Times: How A.I. Killed Student Writing (and Revived It) (Apr. 30, 2026)
  Teachers are moving essays into class, making students write by hand, and using locked browsers to curb generative A.I., while favoring personal, process-focused prompts A.I. can’t easily mimic.
- X (formerly Twitter): LLM Knowledge Bases (Apr. 2, 2026)
  LLMs can compile source documents into a markdown wiki, maintain summaries, backlinks, and categories, and handle ingestion, linting, and improvement. Using Obsidian as an IDE enables Q&A, markdown or slide outputs, visualizations, and easy filing back into the wiki.
- WSJ: Sam Altman Attack Suspect Charged With Attempted Murder (Apr. 13, 2026)
  A 20-year-old, Daniel Moreno-Gama, allegedly threw a Molotov-style device at Sam Altman’s San Francisco home, then attacked OpenAI’s headquarters, carrying an anti-AI manifesto with names and addresses.
- Inside Higher Ed: 4 in 10 Students Say AI Will Influence Their Career Choice (Apr. 30, 2026)
  Nearly half of college-eligible students say AI will influence their career choice, and 10% changed majors. Many feel uncertain, anxious, and depressed about job prospects.
- NY Times: Is A.I. a Threat to Humanity? Not in This Trial. (Apr. 30, 2026)
  Elon Musk is suing OpenAI for breaching its founding contract, seeking $150 billion, unwinding its for-profit shift, and Sam Altman’s removal. Judge barred arguments about A.I. causing human extinction, and lawyers focused on contract facts, credibility, and timing.
- Lawframe: AI Companies Can’t Regulate Themselves. They Should Regulate Each Other. (Apr. 29, 2026)
  Competition undermines AI safety, creating a race to the bottom, information gaps, rapid change, and irreversible risks. A supervised self-regulatory organization, modeled on finance SROs and centered on the Frontier Model Forum, could mandate membership, set rules, and enforce oversight.
May 7, 2026
AI Industry Expansion Meets Growing Security Risks (Links) – May 6, 2026
- mistral.ai: Remote agents in Vibe. Powered by Mistral Medium 3.5 (Apr. 29, 2026)
  Mistral released Mistral Medium 3.5, a 128B model with open weights that powers cloud-based, async coding agents in Mistral Vibe and Le Chat.
- WSJ: With Mythos, AI Pays for Itself (Apr. 14, 2026)
  Anthropic kept Mythos private, reported thousands of software vulnerabilities to authorities, and avoided a public release that could enable abuse.
- WSJ: Why Amazon Is Buying Starlink Rival Globalstar in $11 Billion Deal (Apr. 14, 2026)
  Amazon is buying Globalstar for about $10.8 billion to gain spectrum, satellites, and infrastructure, boost its Leo phone-to-satellite service planned for 2028, and rival SpaceX’s Starlink.
- Bloomberg: OpenAI Share Demand Drops on Secondary Market as Anthropic Gets Investors (Apr. 1, 2026)
  OpenAI shares are dropping on secondary markets, with institutional sellers struggling to find buyers for roughly $600 million in stock. Investors are pouring capital into Anthropic.
- CNBC: Struggling shoe retailer Allbirds makes bizarre pivot from shoes to AI, stock explodes more than 300% (Apr. 15, 2026)
  Allbirds announced a pivot from shoes to AI compute infrastructure as NewBird AI, seeking up to $50 million, and its shares jumped over 300%.
- ZDNET: How AI has suddenly become much more useful to open-source developers (Mar. 31, 2026)
  AI coding tools have recently improved, and maintainers are using them to fix, modernize, and revive many single-maintainer open-source projects.
- Matija Franklin: AI Agent Traps (Mar. 31, 2026)
  AI agents inherit LLM vulnerabilities, but their autonomy, persistence, and tool access create a new information-environment attack surface. Web pages, emails, APIs, and databases can be weaponized.
- NY Times: Why It’s Crucial We Understand How A.I. ‘Thinks’ (Apr. 15, 2026)
  Modern giant neural networks achieve remarkable abilities, but operate as opaque black boxes whose internal rules we can’t easily understand.
- NY Times Opinion: Don’t Use A.I. to Do This (Apr. 15, 2026)
  A writer embraces A.I., uses it constantly, and celebrates its everyday conveniences via a cheeky digital agent called the Gooch. He warns against using generative A.I. for art.
- WSJ: To Lure Top AI Talent, Startups Are Turning to Cold Hard Cash (Mar. 30, 2026)
  AI startups, flush with venture capital, are offering much higher cash pay, larger bonuses, and creative incentives, with median base engineering offers rising from $160k to $200k.
- WSJ: Why the Cost of Your Coffee Has Soared—and Isn’t Going Down Soon (Mar. 29, 2026)
  Reverie Roasters raised its Boneshaker Espresso price from $15 to $17, and will raise it to $18, as tariffs, crop failures, higher rents, and labor costs squeeze margins. Coffee futures and geopolitical worries are pushing commodity prices up again.
May 6, 2026
Blog Articles: May 5, 2026
- Box Blog: Claude Opus 4.7 delivers powerful performance with higher efficiency vs Opus 4.6 (Apr. 16, 2026)
  Claude Opus 4.7 delivers significantly higher efficiency than Opus 4.6, requiring fewer LLM and tool calls, lower latency, and 30% less AI Unit usage.
- NY Times: Musk Says He ‘Was a Fool’ to Provide OpenAI’s Early Funding (Apr. 29, 2026)
  Elon Musk said he regretted funding OpenAI, calling himself “a fool” and accusing it of breaching its founding agreement by partnering with Microsoft and commercializing its work.
- WSJ: AI Worries Have Returned to Wall Street. Now Come Earnings. (Apr. 28, 2026)
  Shares of companies tied to OpenAI, including Oracle, CoreWeave, and SoftBank, slid after reports that OpenAI missed revenue and user targets, reviving worries about AI investments, profits, and circular financing.
- WSJ: The Clock Is Ticking for Big Tech to Make AI Pay (Apr. 30, 2026)
  Microsoft, Amazon, Meta, and Alphabet are pouring record capital into AI, driving $133 billion in Q1 spending and projected $725 billion this year.
- WSJ: Can We Handle the Magnificent 10? (Apr. 29, 2026)
  Anticipated IPOs from OpenAI, Anthropic, and xAI could add trillions, but risky, often-unprofitable AI finances worry markets.
- WSJ: Shark Tank’s Mr. Wonderful Is Planning One of America’s Biggest Data Centers (Apr. 29, 2026)
  Kevin O’Leary is leading a 40,000‑acre Utah data‑center project, expected to need as much as 9 GW (😱) and on‑site natural‑gas power.
- WSJ: White House Opposes Anthropic’s Plan to Expand Access to Mythos Model (Apr. 29, 2026)
  The White House blocked Anthropic’s plan to give 70 more groups access to the Mythos AI, citing national security, cybersecurity, and computing power concerns.
- The Verge: Google and Pentagon reportedly agree on deal for ‘any lawful’ use of AI (Apr. 28, 2026)
  Google signed a classified deal letting the US Department of Defense use its AI for any lawful purpose, despite employee protests. The contract bars domestic mass surveillance, prohibits autonomous weapons without human oversight, and gives Google no veto.
- WSJ: Why AI Startup Offices in NYC Are Flashy but Mostly Empty (Apr. 28, 2026)
  Cash-rich AI startups are leasing large, high-end Manhattan offices, often with many vacant desks, to signal credibility, secure growth room, and enjoy better workspaces.
- WSJ: Ex-Twitter CEO’s AI Startup Raises Funds at $2 Billion Valuation (Apr. 28, 2026)
  Parag Agrawal’s Parallel Web Systems raised $100 million in a Sequoia-led Series B, valuing it at $2 billion. Its platform enables autonomous AI agents to search the web for enterprise tasks, and it will expand sales, marketing, and research.
- The Times: Two-thirds of babies watch screens — some for eight hours a day (Apr. 28, 2026)
  A report finds two-thirds of babies under two use screens, some for up to eight hours a day. Researchers link early screen use to sleep, behaviour, eyesight, and obesity problems.
May 5, 2026
AI safety threats and economic disruption (Links) – May 4, 2026
- NY Times Opinion: After Mythos, Nobody Is Safe From Cybersecurity Threats (Apr. 28, 2026)
  New AI like Claude Mythos can find unknown software flaws, increasing risk for individuals, small groups, and nonprofits. Use unique passwords or passkeys, a password manager, authenticator apps, and keep software updated; companies and government must coordinate protections.
- Cloudflare: Browser Run: give your agents a browser (Apr. 15, 2026)
  Cloudflare renamed Browser Rendering to Browser Run, providing full browser sessions on its global network for AI agents, with CDP and WebMCP support. New features include Live View, human-in-the-loop, session recordings, and higher concurrency for scalable agent automation.
- WSJ: AI Has Made Memory Chips One of the World’s Most Profitable Products (Apr. 30, 2026)
  AI demand has sent memory-chip prices and profits soaring. Samsung, SK Hynix, and Micron posted record results—Samsung made about $30B in one quarter—and they’re forecast to earn roughly $350B in 2026.
- NY Times Opinion: Silicon Valley Is Bracing for a Permanent Underclass (Apr. 30, 2026)
  Silicon Valley fears advanced A.I. could displace millions, concentrate wealth, and create a permanent underclass, as companies race to automate jobs, benchmark models, and cut staff.
- NY Times Opinion: Why Are We Still Driving? (Apr. 30, 2026)
  Self-driving cars promise far fewer deaths, reclaimed hours, and city robotaxi fleets becoming common within a decade or two. Waymo’s sensor-heavy, costly approach and Tesla’s camera-only bet compete.
- NY Times: How Good Is Google’s Gemini AI at Making Travel Plans? (Apr. 30, 2026)
  Google’s Gemini, paired with Ask Maps, sped trip planning by creating itineraries, packing lists, and tailored flight and hotel suggestions using personal data.
- NY Times: A.I. Bots Told Scientists How to Make Biological Weapons (Apr. 29, 2026)
  A.I. chatbots produced step-by-step instructions for making, modifying, and releasing biological agents, and suggested ways to evade detection.
- NY Times: Elon Musk, Testifying at OpenAI Trial, Accuses Executives of Stealing a ‘Charity’ (Apr. 28, 2026)
  Elon Musk sued Sam Altman, alleging OpenAI stole a charity when it shifted to profit, seeking $150 billion, and an unwind. The jury’s verdict could reshape A.I.
- NY Times: A.I. Helps Online Ad Businesses Boom (Apr. 29, 2026)
  AI tools from Google and Meta automate ad creation, targeting, bidding, and measurement, making ads cheaper, more effective, and easier for small businesses to run.
- NY Times Opinion: Rich People Didn’t Used to Look Like This (Apr. 30, 2026)
  Wealthy people now signal status with extreme cosmetic facial work, turning altered faces into luxury badges. Social media, celebrity culture, and influencers have normalized procedures, driving demand across classes, despite health risks and the waning appeal of designer goods.
- University of Cincinatti: The Collapse of Teen Fertility in the Digital Era (Apr. 25, 2026)
  Teen fertility collapsed globally around 2007 as smartphones shifted teens from in-person to phone-mediated peer time, reducing unstructured contact and conceptions. Cross-country timing,
May 4, 2026
OpenAI Crisis Meets Global AI Backlash (Links) – May 3, 2026
- NY Times: What Elon Musk’s Clash With Sam Altman of OpenAI Is Really About (Apr. 28, 2026)
  Elon Musk’s lawsuit against Sam Altman and OpenAI frames a Silicon Valley fight over profit, power, and broken nonprofit promises, with a trial that exposed internal emails.
- NY Times: Is OpenAI Falling Further Behind in the A.I. Race? (Apr. 28, 2026)
  OpenAI missed user and revenue targets, raising doubts about its spending, market position, and IPO.
- Simon Willison: Speech translation in Google Meet is now rolling out to mobile devices (Apr. 27, 2026)
  Google Meet’s speech translation is rolling out to mobile, translating and replaying speech in the speaker’s voice. It supports English, Spanish, French, German, Portuguese, and Italian, but is still alpha.
- WSJ: Meta Is Preparing to Have to Undo Its Manus Acquisition After China Ban (Apr. 27, 2026)
- WSJ: OpenAI Misses Key Revenue, User Targets in High-Stakes Sprint Toward IPO (Apr. 27, 2026)
  OpenAI missed user and revenue targets, raising concerns about its heavy data-center commitments and ability to pay future compute contracts.
- OpenAI: Our principles | OpenAI (Apr. 27, 2026)
  Mission: ensure AGI benefits all humanity by putting powerful, general AI into many people’s hands rather than concentrating control.
- Simon Willison: Tracking the history of the now-deceased OpenAI Microsoft AGI clause (Apr. 27, 2026)
  Microsoft and OpenAI renegotiated their deal, ending the AGI-triggered transfer of Microsoft’s exclusive IP rights. Microsoft keeps a non-exclusive license through 2032, OpenAI stopped paying revenue share, and payments to Microsoft continue through 2030 regardless of progress.
- WSJ: Chip Startup Aims to Shatter AI’s Dreaded Memory Wall (Apr. 28, 2026)
  Majestic Labs, started by ex-Google and Meta chip leaders, built Prometheus servers with AIU chips and up to 128 TB of memory to run trillion-parameter models.
- NY Times: Your Doctor’s Notes Might Be Written by an A.I. Algorithm. Here’s What to Know. (Apr. 28, 2026)
  A.I. scribes record doctor visits, create draft notes, and can reduce paperwork and burnout. But they raise privacy, consent, and accuracy concerns.
- Transformer: The AI safety movement needs normies (Apr. 27, 2026)
  Fears about jobs, surveillance, and extinction mean safety advocates must engage the public, build coalitions, and demand political action.
- WSJ: China’s AI ‘Hotel California’ – WSJ (Apr. 27, 2026)
  Beijing blocked Meta’s $2 billion buy of Manus, likely on national-security grounds, after Manus moved to Singapore for foreign capital. Founders were detained, exits were blocked, and Xi aims to keep AI talent within state-owned, state-linked, or party-backed firms.
- NY Times: From Indiana to Idaho, a Backlash Against A.I. Gathers Momentum (Apr. 27, 2026)
  A growing backlash against A.I. is sweeping the U.S., uniting parents, religious leaders, environmentalists, and former Tea Party activists who fear Big Tech will profit while Americans bear the costs.
- NY Times: Elon Musk and Sam Altman Bring OpenAI Trial Spectacle to Oakland (Apr. 27, 2026)
  Elon Musk sues OpenAI and Sam Altman in Oakland, seeking $150 billion and a return to nonprofit, with Musk, Altman, and tech leaders set to testify.
- WSJ: AI Investment Boosted Economic Growth, While Consumers Tapped the Brakes (Apr. 30, 2026)
  U.S. GDP grew 2% annualized in Q1, below forecasts, as strong business investment, especially in AI, offset softer consumer spending.
May 3, 2026
AI Security Risks and Industry Shakeup (Links) – May 2, 2026
- Ben Thompson: An Interview with Google Cloud CEO Thomas Kurian About the Agentic Moment (Apr. 23, 2026)
  Thomas Kurian outlines Google Cloud’s push for agentic AI with Gemini, secure infrastructure, and deep integration with Google systems.
- WSJ: Microsoft Needs Copilot to Get Back in the Air (Apr. 23, 2026)
  Microsoft’s stock plunged after AI enthusiasm cooled, as Copilot adoption lags and Azure growth slows amid data-center capacity limits.
- SpaceX: SpaceXAI and Cursor (Apr. 23, 2026)
- Simon Willison: Is Claude Code going to cost $100/month? Probably not—it’s all very confusing (Apr. 22, 2026)
  Anthropic briefly moved Claude Code off the $20/month Pro plan to $100–$200/month Max plans on its pricing page, then reverted the change after public outcry. They said it was a small test, but it caused confusion, anger, and lost trust.
- WSJ: Anthropic Probes Possible Unauthorized Access to Mythos AI Model (Apr. 22, 2026)
  Anthropic is probing possible unauthorized access to Mythos via a contractor, stoking fears it could enable cyberattacks.
- NY Times: Anthropic’s New Mythos A.I. Model Sets Off Global Alarms (Apr. 22, 2026)
  Anthropic’s new AI model, Mythos, can find and exploit hidden software flaws, prompting emergency responses from central banks, intelligence agencies, and governments worldwide.
- Martin Alderson: Figma's woes compound with Claude Design (Apr. 18, 2026)
  Regulators in Australia, South Korea, and Singapore are monitoring Anthropic’s Mythos, fearing its advanced coding could expose cybersecurity flaws and threaten banks.
- WSJ: Anthropic CEO Dario Amodei Lands White House Meeting as Feud Thaws (Apr. 17, 2026)
  Dario Amodei will meet White House chief of staff Susie Wiles as Anthropic previews its Mythos AI, briefing agencies, and offers advance access to reduce cyber risks.
- NY Times: How Do You Measure an A.I. Boom? (Apr. 17, 2026)
  METR’s time-horizon chart shows A.I. agents’ task-length doubling from seven months to about three to four months, fueling massive investment, hype, and debate.
- NY Times: Want to Speak to the Manager? At a New San Francisco Store, That’s A.I. (Apr. 21, 2026)
  Andon Market in San Francisco is billed as the first retail boutique run by an A.I. agent, Luna, who picked odd inventory, ordered many candles, and set quirky prices.
- NY Times: Leaked Code for Anthropic’s Claude Code Tests Copyright Challenges in A.I. Era (Apr. 22, 2026)
  A leaked Anthropic source code was rewritten by a student using A.I. agents, shared online, and left online despite takedowns for copyright.
- Reuters: Asia regulators monitor Anthropic's Mythos for potential banking risks (Apr. 20, 2026)
  Regulators in Australia, South Korea, and Singapore are monitoring Anthropic’s Mythos, fearing its advanced coding could expose cybersecurity flaws and threaten banks.
- TechCrunch: Stanford report highlights growing disconnect between AI insiders and everyone else (Apr. 13, 2026)
  Stanford’s report finds a growing gap between AI experts, who expect net benefits, and the public, which fears job losses, higher costs, and health impacts.
- NY Times: He Warned About the Dangers of A.I. If Only His Father Had Listened. (Apr. 13, 2026)
  Ben Riley, a skeptical writer about A.I., discovered his father, Joe, had used A.I. research to refuse leukemia treatment, despite doctors’ warnings.
- The Chronicle of Higher Education: Teaching: The students who won’t use AI (Apr. 16, 2026)
  Many students use AI for coursework, but a sizable minority avoid or rarely use it, and peers sometimes shame AI users. Gen Z’s views grew more negative—less excitement, less hope, more anger, steady anxiety.
- WSJ: Has the Era of the Mega-Layoff Arrived? (Apr. 15, 2026)
  Major tech and corporate employers are cutting vast swaths of staff, sometimes 30–40%, as investors reward deep job cuts, and companies cite AI, high AI costs, and pandemic overhiring.
- WSJ: The U.S. Has Long Been a Nation of Inventors—and Luddites (Apr. 22, 2026)
  America has long embraced invention, yet repeatedly pushed back against innovations seen as threats to autonomy, health, or community.
- WSJ Opinion: The Biggest AI Risk Is Foolish, Fear-Driven Policies (Apr. 21, 2026)
  AI’s impact has been overstated, with modest productivity gains predicted, only limited task automation, and no widespread job collapse.
- The Free Press: I Thought I Was Autistic. I Was Wrong. (Apr. 24, 2026)
  Growing research about a “female autism phenotype” has led to sharply rising diagnoses in young women, but experts now warn the spectrum may have become so broad it risks losing clinical meaning.
May 2, 2026