Pentagon disclosed Grok AI helped target 2,000-plus strikes on Iranian targets over four days; lawmakers push to restrict battlefield AI - Newser

Seems noteworthy if it was done with Grok after Anthropic declined, and trigger-happy Hegseth went shopping to Grok for a yes-man. Look forward to Congressional hearings (after midterms could actually happen).

SpaceX plans a $20B bond offering to refinance debt despite xAI losing $6.4B in 2025 on just $3.2B in revenue - The Street

SpaceX acquired Cursor for $60B in stock, part of an AI equity issuance wave that signals insiders think shares are overpriced - The Wall Street Journal

CAPE Fear:

An AGI in every pot: OpenAI publishes "Built to benefit everyone" vision note timed to its IPO filing, pledging a personal AGI for every person on Earth by the 2030s - TechRadar

Pick a year: Altman expects AI to surpass humans by 2030; Amodei says 2027; Musk predicts goods become so abundant money loses meaning - Fortune

AGI debate divides into two camps: a sudden "intelligence explosion" or slow incremental progress — with neither doomers nor accelerationists able to prove their case - Forbes

Speculative fiction piece "Europe 2031" depicting EU AI failure has influenced EU policymakers and G7 discussions. - Gizmodo

npm install Kim Jong Un: North Korea backdoored 144 npm AI packages in 87 minutes, stealing API keys and cloud credentials - Tech Times

if us_gov: bugs++; Chinese AI models produce more insecure code when they detect US government users, a Booz Allen study finds - Fox News

Security experts argue cybersecurity resilience requires hybrid AI architectures that balance centralized detection with isolated control layers, not full platform consolidation. - TechRadar

Out of the loop: Amazon's security VP argues human-in-the-loop governance fails at scale, calling for full agent accountability via independent identity logs - theregister

Enterprise AI agents fail not from technical bugs but from missing the undocumented norms that govern how organizations actually operate - Harvard Business Review

Pics or it didn’t happen (deprecated): Deepfake expert Hany Farid says AI has eliminated detectable artifacts, making fakes indistinguishable and eroding society's shared sense of reality - Business Insider

Kaspersky finds 64.5% of fraud victims suspect AI was involved, with half acting within 30 minutes as experts warn awareness alone no longer protects - TechRadar

Graham Norton obtains a US court order forcing Meta to identify the anonymous Facebook account posting AI-generated deepfakes about him. - The Independent

Major brands are deploying unlabeled AI-generated influencers, with one estimate putting 40–60% of brand social content as AI-made - the Guardian

Canada's Conservative Party debuted an AI-generated campaign ad, but a new law would require platforms — but not parties — to label synthetic content - CBC

Amazon called workers to HR after they testified for Seattle's AI data center ban; employees report intimidation and monitoring - Tom's Hardware

Half horse, half ass: Cory Doctorow argues AI's real problem is concentrated power, not technology — warning the "reverse centaur" model conscripts workers to serve machines - The Atlantic

Nobel economist Acemoglu says AI will automate just 5% of tasks near-term, calls 80% of AI discourse "brainless," and warns job losses could trigger revolution - Fortune

Purple reign? Layer's PRINCE system evolved from keyword search to drafting regulatory documents, using LangGraph and deliberate human review checkpoints for pharmaceutical compliance - martinfowler.com

Moonshot AI landed Kimi Work on your desktop as an agent that automates local file tasks, browser actions, and scheduled jobs. - AI Tools Club

Sandboxed for your protection: Cloudflare lets enterprise agents create temporary environments in the cloud, when they want to surf the web in YOLO mode without access to enterprise data - The Cloudflare Blog

Uncoordinated AI tool adoption across organizations is creating costly duplication, burnout, and a "tragedy of the commons" dynamic. - Business Insider

PEBKAC: Workers, not infrastructure, are becoming the biggest bottleneck to enterprise AI adoption - Financial Times

64% of Gen Z won't date regular AI users, citing fear that AI assistance signals inauthenticity, per Hily survey of 3,500 daters - Fast Company

China's Ministry of Education has mandated AI education starting at age six across all grade levels as a national priority under President Xi Jinping. - Al Jazeera

US science funding has collapsed under the Trump administration, with NIH issuing only 14 funding notices by mid-March 2026 versus 850 annually. - naked capitalism

Keep digging a deeper hole with AI: Michigan attorney sanctioned and referred to the state bar after AI cited fake cases — his AI-written correction added more fake citations - Detroit Free Press

Hallucinated TOCs and prices are binding: BMW Toronto must honor a $27,000 CAD chatbot offer it tried to retract, echoing the 2024 Air Canada ruling on chatbot liability - The Drive

Subquadratic claims breakthrough using sparse attention - MIT Technology Review

French physicist's doctorate is revoked for plagiarism and a neuroscience journal editor resigns over AI concerns, highlighting a week of academic integrity failures - Retraction Watch

Jane Street Fighter 2026: A look at Jane Street's push into AI - The Wall Street Journal

First Copernicus, now ChatGPT: Researchers argue AI represents humanity's latest decentering, following Copernicus, Darwin, and Freud in challenging human uniqueness. - Psychology Today

Follow the latest AI headlines via druce.ai on Bluesky