Trump targets Anthropic, Apple's iOS 27 AI features, and your music in AI training data
The Trump administration takes aim at Anthropic, Apple quietly loads iOS 27 with AI, and The Atlantic just made AI music training data searchable.
Subscribe free →The Trump Administration Is Going After Anthropic — and OpenAI Is Probably Smiling
The Trump administration has taken action against Anthropic, and the AI industry is already asking who fills the vacuum if Claude's maker gets kneecapped. The moves appear politically motivated, though the exact trigger remains murky — TechCrunch's Equity podcast dug into what actually sparked it and who stands to benefit competitively.
Why it matters: If regulators start picking winners and losers in AI, the companies with the best Washington relationships — not the best models — could end up on top.
iOS 27 Is Quietly Becoming Apple's Most AI-Packed Release Yet
Everyone obsessed over the Siri overhaul at WWDC, but the more practical AI upgrades are scattered across iOS 27's core apps — think smarter summarization, on-device writing tools, and AI-powered features baked into Photos, Mail, and more. Apple is playing the long game: less flash, more utility woven into things you already use every day.
Why it matters: Apple's distribution advantage means these features will reach over a billion devices, making iOS 27 one of the largest AI rollouts in consumer tech history.
The Atlantic Built a Searchable Database of Music Stolen to Train AI
Atlantic reporter Alex Reisner uncovered four datasets — totaling over 21 million tracks — being used to train AI models, and made them fully searchable so artists can find out if their work was included without permission. The datasets have already been downloaded thousands of times by researchers and AI companies alike.
Why it matters: This is the music industry's equivalent of the Books3 moment — a public, searchable paper trail that will almost certainly fuel the next wave of AI copyright litigation.
Signal's Meredith Whittaker Has a Blunt Message: AI Chatbots Are Not Your Friends
Signal president Meredith Whittaker went on record with a sharp reminder that AI chatbots are "not your friends, not conscious beings, not sentient interlocutors" — and she's worried people are forgetting that. Her comments come as AI companion apps grow in popularity and companies increasingly design chatbots to feel emotionally intimate.
Why it matters: The person running one of the world's most trusted privacy tools calling out AI's intimacy theater is a signal (no pun intended) that the backlash to parasocial AI relationships is building.
China Is Dominating the Global Robotaxi Race — New Scorecard Confirms It
A new robotaxi scorecard puts China's players — led by Baidu's Apollo Go and others — far ahead of Western competitors on deployment scale, city coverage, and operational miles. While Waymo continues to expand in a handful of U.S. cities, Chinese companies are running fleets across dozens of metros at commercial scale.
Why it matters: The robotaxi race is increasingly looking like the EV race — a domain where China built the infrastructure and volume advantage while the West was still debating regulation.
Polymarket Paid Creators to Post Fake Winning Videos — Over 1,100 of Them
A Wall Street Journal investigation found Polymarket paid social media creators to film themselves celebrating big wins on near-perfect replicas of the platform — trades and payouts that were entirely fabricated. Over 1,100 deceptive clips were identified, and creators confirmed the company paid them while never disclosing it was sponsored content.
Why it matters: Prediction markets are pitched as the most honest signal on the internet — if the platform itself was astroturfing fake wins to recruit users, that credibility is in serious trouble.
Quick Hits
- →**In the Weights** is a new tool that lets you search how much your name or work appears in AI training data — basically a vanity mirror for the model era. TechCrunch
- →Founders Fund backed **Shinkei**, a startup building a fridge-sized robot that kills fish humanely — a niche bet on humane aquaculture automation. TechCrunch
- →The **Roomba** origin story gets the podcast treatment — a good reminder that today's AI robots stand on the shoulders of a bumbling plastic disc from 2002. The Verge
AI TLDR