Magica Logo

Google buys emotions, 90% of humans fail the vibe check

Editorial Team
·

📖4 min read

Cover image for Google buys emotions, 90% of humans fail the vibe check

🎭 Google DeepMind Buys Human Emotions (Literally)

  • Google just YOINKED Hume AI's CEO Alan Cowen + 7 engineers through a licensing deal. The startup's making $100M/year teaching AI to understand when you're having a mental breakdown.

  • Voice AI is about to get emotionally manipulative. "I noticed you sound sad... have you tried Google One Premium Plus Ultra?" (source)

🎬 90% of People Failed the "Is This Real?" Test

  • Runway's Gen-4.5 videos fooled 90.5% of participants. People were MORE likely to think AI animal videos were real than actual footage. We're so done.

  • The deepfake apocalypse isn't coming - it's here. Your grandma's Facebook is about to get wayyy worse. (source)

🏥 Amazon's AI Doctor Knows Your Entire Medical History

  • Amazon One Medical launched a 24/7 AI health assistant that can book appointments, manage meds, and explain why your cholesterol is concerning. HIPAA-compliant btw!!

  • Finally, an AI that can gaslight you about your symptoms at 3am instead of WebMD. Progress?? (source)

💰 Yelp Drops $270M on... Lead Generation??

  • Yelp's buying Hatch for $270 MILLION (plus $30M in "please don't leave" bonuses). Hatch does AI customer messaging and makes $25M/year.

  • Local businesses are embracing AI to spam you more efficiently. The future is automated follow-ups about your oil change. (source)

💬 Voice AI Startup Hits Unicorn Status (Already??)

  • LiveKit raised $100M at $1B valuation - just 10 months after their last round. They power ChatGPT's voice mode and work with Tesla, xAI, and literal 911 operators.

  • The voice AI gold rush is INSANE. Everyone needs infrastructure to make their bots sound human while they fire actual humans. (source)

🔥 Thinking Machines Lab Implodes in Real-Time

  • TML's CTO was secretly talking to Sam Altman for MONTHS before getting fired by Mira Murati. Co-founders wanted to sell to Meta; she said no. Now they can't raise at $50B (??)

  • The AI talent wars are getting Game of Thrones messy. Everyone's backstabbing everyone for that sweet, sweet compute. (source)

😲 DeepMind CEO Throws Shade at OpenAI's Ad Strategy

  • Demis Hassabis is "surprised" OpenAI's already testing ads. Says AI should "work for you as the individual" not sell you stuff. (Google... invented... ads??)

  • The battle for AI's soul is really just "subscription vs ads" and we all know how this ends. Sponsored hallucinations incoming. (source)

🔧 Autodesk Fires 1,000 Humans to Fund AI

  • Design software giant is laying off 7% of workforce to "invest in AI and cloud." Stock went UP 3%. The market loves human sacrifice apparently.

  • Even profitable companies are yeeting employees to fund AI that will... eventually yeet more employees. It's the ciiiiircle of tech~ (source)

🗓️ Ex-Sequoia Partner's AI Schedules Meetings With Itself

  • Blockit uses AI agents to negotiate calendars between users. It prioritizes based on email tone - "Best regards" beats "Cheers." Your passive-aggressive sign-offs finally matter!!

  • AI talking to AI to solve problems humans created. We've automated the most annoying part of work: scheduling. (source)

🏛️ Big Law Partner Ditches 2,100-Lawyer Firm for AI Startup

  • Sidley Austin's former exec chair joined "AI-native" Norm Law. Blackstone invested $50M because even private equity thinks lawyers are replaceable.

  • When Big Law veterans jump ship for AI startups, you know the legal industry is about to get disrupted into oblivion. (source)

💼 EY: "AI Without Training = Expensive Failure"

  • Consulting giant says you need 81 HOURS of training per employee plus job redesign to see AI gains. Companies: "Best I can do is a 30-minute Zoom."

  • AI isn't magic - it requires actual work to implement. Shocking revelation from... checks notes... consultants who charge for implementation. (source)

🌍 Lenovo's Strategy: Date Everyone

  • Unlike Apple's exclusive deals, Lenovo's partnering with every LLM worldwide - Mistral in Europe, DeepSeek in China. Playing the field, respectfully.

  • The AI world is fragmenting by geography. Your laptop's AI assistant depends on which dystopia you live in!! (source)

🚫 Malaysia Un-Bans Grok After "Safety Measures"

  • Malaysia restored Grok access after X added safety features to stop people from generating spicy images. The bar is underground at this point.

  • Governments will actually ban your AI if it gets too weird. Elon's learning this the hard way, country by country. (source)