Good morning. Remember when your parents told you video games would rot your brain? Well, the joke’s on them because catching Pikachu is now the ultimate test for artificial intelligence. Researchers are streaming “Claude Plays Pokémon” on Twitch and using it like a live benchmark. Apparently the key to smarter AI isn’t complex math puzzles, it’s beating gym masters and catching Snorlax. Goals.
Wikipedia accidentally created a cheat sheet for AI detectors. Editors spent 2 years obsessively cataloging awkward cliches and analytical fluff chatbots love to spew. Tech entrepreneur Siqi Chen turned their detailed list into Humanizer, a plugin that instantly earned over 1,600 stars on GitHub. Humanizer teaches chatbots to sound human and casual using Wikipedia’s own rules. The irony? Once bots perfectly blend in with human editors, Wikipedia’s whole “sum of human knowledge” promise becomes meaningless. Welcome to Grokipedia. (Ars Technica)
Meta is trying to buy Congress. In 2025, Meta spent over $26 million lobbying lawmakers, nearly 2x what Amazon, Google, or Apple shelled out. In just the last quarter they dropped $6.5 million. Meta now has 1 lobbyist for every 6 members of Congress, pouring cash into shaping AI regulations and kids’ online safety laws. While Nvidia and other AI companies ease off, Meta is spending aggressively to control exactly how the rules are written. (Axios)
Wispr Flow helps teams work 4x faster in every app. Speak prompts, dictate PRDs, generate code, and reply to messages using your voice. Flow precisely captures your ideas at the speed you think.
Used by teams at Clay, OpenAI, and Amazon to execute 20% faster. Because it works, magically.
Stop typing. Start talking to your apps. Try Wispr Flow for free.
800 creators have accused AI of grand theft artistry. Scarlett Johansson, REM, and hundreds of artists have joined the “Stealing Isn’t Innovation” campaign to stop AI slop. They’re calling out AI companies for ripping off their work to train generative AI models, demanding control and compensation. Industry orgs say licensing and partnerships are the answer, a deal that ensures the big players at least get paid. Individual creators may be left wondering if their human creativity still matters. (The Verge)
Jeff Bezos is building a satellite empire to rival Musk’s Starlink. Blue Origin just unveiled TeraWave, a massive satellite internet network aimed at businesses and governments. TeraWave plans to launch 5,408 satellites by late 2027, delivering speeds up to 6 terabits per second – faster than anything that’s currently available. Ironically, Bezos will be competing with Amazon’s own consumer satellite network, Leo. No doubt he has plans for space data centers too. (BBC)
AI sovereignty is global dependence with better branding. Governments worldwide plan to dump $1.3 trillion into AI infrastructure by 2030, promising full independence. There’s one tiny problem: AI supply chains are global. Chips are US designed and built in Asia. The data comes from everywhere. And the smartest people? They chase innovation wherever it happens. Isolation might make politicians feel powerful, but it doesn’t build AI. Countries like Singapore, Israel, and South Korea know the truth. Winning isn’t about expensive isolation. It’s about picking the right friends. (MIT Review)
Humanoid robots look cool, but can they do anything useful yet? At Davos, robotics experts admitted those flashy parkour bots you see on stage still can’t reliably make coffee or fold your laundry without burning millions of dollars. Musk swears Tesla’s Optimus robots will be everywhere by late 2027. But behind the scenes, most humanoids aren’t autonomous at all. They’re remote controlled by people, a dirty little secret the robotics industry prefers to hide. (Business Insider)
fun stats
🤖 16,000. Humanoid robots went to work in 2025. By 2027, over 100,000 could be on the clock.
💬 94%. B2B buyers use AI during their buying process. Most are aware of the top brands, and use LLMs to compare vendors.
🤓 54%. Quant investors don’t use gen AI for investing. Wall Street’s most data obsessed need clean, structured data and repeatable models – and don’t think AI can beat the market.