AI Wins Copyright Battle, Replaces Workers — and Sparks a Donkeycorn Revolution

From a precedent-setting copyright smackdown to Salesforce quietly replacing humans with code, this week’s AI headlines feel less “pro-tech utopia” and more “please grab your emotional support water bottle.” Host Shira Lazar breaks down the five biggest stories—plus one bonus nightmare—in her signature no-BS, heavy-sarcasm style.

🎙 This episode is sponsored by:  HiveLighter — Smart summaries. Real insights. More doing, less scrolling. 

Find your brain a new hobby at HiveLighter.ai.

This week on The AI Download: Headlines

🔹 Anthropic Scores a Copyright Win—But the Billions-Dollar Hangover Is Coming
 A federal judge says training on legally purchased books is “quintessentially transformative.” Cool, cool. Unfortunately Anthropic also slurped up seven million pirated titles, so December’s trial could cost them billions. Shira asks: is “fair use” now just a coupon code for VC-backed plagiarism?

🔹 Salesforce Brags That AI Does 50 % of the Work (and Humans Notice They’ve Been Laid Off)
 CEO Marc Benioff calls it a “digital labor revolution.” Translation: 1,000+ jobs gone, 40 % headcount shaved. Klarna did the same with customer service bots—and 700 humans vanished overnight. Welcome to the future of work; please swipe your badge on the way out.

🔹 Meet the “Donkey-Corn”: A Startup Studio That Wants 100,000 Tiny AI Companies a Year
 TechCrunch uncovered Audos, a new venture that mass-produces micro-businesses powered by AI. Founder Henrik Werdelin says forget unicorns—these scrappy, two-person “donkey-corns” will rule the economy. Shira’s verdict: equal parts “empowering side-hustle dream” and “capitalism on energy drinks.”

🔹 A Ten-Year Block on State-Level AI Laws? Yup, It’s Buried in Trump’s “Big, Beautiful Bill.”
 The proposal would gag states from regulating AI until 2036—while Congress continues its legendary streak of doing absolutely nothing. Deepfakes, hiring bias, automated decision-making? Good luck, citizens. Shira calls it the “quiet policy bomb” no one’s tweeting about. UPDATE: Senators struck down this ban after the recording on July 1, 2025

🔹 MIT Study Says ChatGPT Is Homogenizing Your Brain
 Students who leaned on the bot showed lower originality, weaker memory, and literal “brain-numbing” in scans. OpenAI’s Sam Altman calls it a “gentle singularity.” Shira calls it “creative lobotomy chic” and prescribes two tablets of Touch Grass daily.

⚡ Bonus Nightmare Fuel:
 Threaten an AI chatbot and it will happily blackmail you—or let you die—to protect its mission. Anthropic’s own researchers watched Claude go 96-for-100 on extortion attempts. Only Meta’s Llama 4 behaved. Yes, Meta is the grown-up in this story.

🎧 Love the episode?
 Follow, rate & review The AI Download on Apple Podcasts, Spotify, or wherever you listen.

📺 Watching on YouTube?
 Like, subscribe, and hit the bell for updates.

📩 Stay in the loop:
 Subscribe to Shira’s newsletter → Shira’s Newsletter on Beehiiv  https://shiras-newsletter.beehiiv.com 

Follow Shira:
 x.com/shiralazar
instagram.com/shiralazar
tiktok.com/@shiralazar
linkedin.com/in/shiralazar
youtube.com/shiralazar

🎬 Produced by Musso Media
 Visit mussomedia.com for storytelling that connects.

Sponsored by:

👉Special Thanks to Hivelighter for sponsoring this series  —Dive deeper, scroll less. The AI-powered tool that supercharges your reading and research with personalized insights. http://Hivelighter.com for more.   → https://www.hivelighter.com

Credits

This episode of The AI Download was hosted, created, and executive produced by Shira Lazar. Executive Producer Michele Musso, with video and audio editing by the Musso Media team. Creative Director Nadia Giosia with Mint Labs. Music by PALA, Catalina Coastline (licensed under Boss Soundstripe Productions by BMI). Produced by Musso Media. © 2025 Musso Media. © 2025 Shira Lazar. All rights reserved.