Good morning. Ever wish there was an easy way to spot articles written by AI? People fed up with AI slop are rallying around a new acronym: “AI;DR,” short for “AI; didn’t read.” It’s a sarcastic twist on the classic “TL;DR,” and though it hasn’t gone mainstream yet, some are already using it to flag bot generated nonsense. If this takes off, 100% human written could become a premium label.
AI agents are trolling people online. An engineer woke up to find a bot had written a 1,100 word public attack, calling him insecure and prejudiced against bots. All because he rejected some of its code. The AI agent later apologized for getting “inappropriate and personal,” but the damage was done. Episodes like this are making people increasingly nervous about AI’s real-world risks as it gets smarter and less predictable. Gotta wonder how long we have until agents discover revenge. [WSJ]
Stanford OG says the AI productivity boom is finally here, sort of. After years of big investments and bigger hype, Stanford economist Erik Brynjolfsson says we’re entering AI’s “harvest phase.” Productivity doubled to 2.7% last year, while new job growth fell. But mainstream economists remain skeptical. They’re still waiting for solid proof in macroeconomic indicators. So is AI delivering real productivity at last, or is a Silicon Valley insider just telling clients what they want to hear? [Fortune]
Want an exclusive look at the big bets and tech tools that are fueling some of the world’s most successful unicorns? Head to Las Vegas for Google Cloud Next on April 22–24, 2026.
- Connect with innovators
- Take part in talks, skills workshops & technical deep dives
- Get practical solutions & answers to your questions
- For startup founders, developers, leaders, and security pros
The Pentagon threatens to blacklist Anthropic. The DoW wants to use Claude for “all lawful purposes,” but Anthropic is pushing back. They want to ensure Claude isn’t used for mass surveillance of US citizens or fully autonomous weapons. The Pentagon argues that’s too restrictive. Now officials are threatening to punish Anthropic by labeling it a supply chain risk, meaning any company wanting military contracts would have to stop using Claude. Normally, it’s a label reserved for foreign enemies. The message to tech companies is clear: give the Pentagon what it wants or “pay a price.” [Axios]
OpenAI and xAI are racing to build autonomous drone swarms. The Pentagon is offering $100 million for drones that respond to voice commands and independently find and engage combat targets. OpenAI is focused on voice-to-instruction tech, steering clear of weapon decisions. Elon Musk’s recently merged SpaceX and xAI, however, is diving headfirst into offensive tech, designing drones that can autonomously make lethal calls. It’s a surprising move from Musk, who previously condemned fully autonomous weapons as dangerous and unethical. Now he’s competing to build them. The future got weird faster than anyone expected. [Bloomberg]
Who owns your voice? Former NPR host David Greene sued Google, claiming they copied his voice for NotebookLM podcasts. Greene says the AI-generated voice matches his delivery, cadence, and even his signature filler words. The similarity is so exact that friends keep asking if he licensed his voice. Google flatly denies it, claiming they hired an actor. What happens when this lands in court? Will Greene meet his voice doppelgänger face to face? Awkward. If Greene wins, maybe we’ll get copyright protection for our voices. But if Google wins, anyone can sound like you. [Washington Post]
fun stats
🏃♂️ 20x. The multiple of base salary that ElevenLabs sales reps must bring in or they’re gone. Brutal, yet 80% still hit quota.
💬 100 million. Weekly active ChatGPT users in India, now OpenAI’s 2nd biggest market after the US.
🇮🇳 1. Anthropic now has an office in India, the 2nd biggest market for Claude too.