Every week there's more AI news than any one person can actually read. So I'm doing the reading for you. Five stories, five days, no fluff. Here's what mattered this week.
Monday: The 24-year-old "AI prophet" just got a real lesson in risk.
Leopold Aschenbrenner became famous in Silicon Valley for a 165-page essay predicting the world was underestimating how fast AI would move. He turned that thesis into a hedge fund, Situational Awareness, and it worked. Spectacularly. At its peak this month, the fund held roughly $45 billion in assets.
Then a selloff in AI-related chip stocks hit, margin calls came due, and by the end of the week his fund had been forced to offload its leveraged positions to Ken Griffin's Citadel at a discount. Assets fell to around $10 billion. The fund is still up significantly for the year, but the unraveling is a reminder that conviction and leverage are two very different things, and that even the person who saw AI coming can get burned by how fast the market around it moves.
Tuesday: Travis Kalanick spent 8 years building in total silence, on purpose.
In a conversation with a16z, the Uber cofounder talked about running his current company, Atoms, in stealth mode for nearly a decade before saying a word publicly. No name on LinkedIn. No press. Just a team solving hard problems in food automation, industrial robotics, and autonomy.
His argument for staying quiet was simple. Fame attracts people who want to be known. Silence attracts people who want to build. If you're hiring for a team of doers instead of a team of personalities, going dark might be the advantage nobody's talking about.
Wednesday: Almost half of new podcasts might not have a human behind them anymore.
The Wall Street Journal dug into a trend that's been building for a while: AI-generated podcasts, sometimes called "podslop," are now a meaningful share of everything new hitting the feeds. Companies like Inception Point AI are reportedly producing thousands of episodes a week, sometimes for around a dollar each, using synthetic hosts that can sound convincingly human if you're not listening closely.
The uncomfortable question isn't really about podcasts. It's about how much of what we consume every day, across every format, we can still confidently call real.
Thursday: China didn't just catch up in AI. It found its own lane.
Tech analyst Grace Shao broke down how Chinese AI companies went from playing catch-up to shipping serious, competitive products, especially in open-source models, robotics, and AI agents. Her read isn't that China copied the US playbook. It's that resource constraints, particularly around chips, forced Chinese labs into different tradeoffs, and some of those tradeoffs turned into real advantages.
If you've been assuming the US has a clean lead in AI, this conversation is worth the watch.
Friday: Someone stole $70 million in bitcoin from a system that was supposedly unhackable.
Nearly 1,000 bitcoin vanished in 41 minutes from a cold storage system, the kind of offline setup that's supposed to be the gold standard for crypto security specifically because it isn't connected to the internet. Investigators are still piecing together how hackers compromised it, but the early read is that the weak point wasn't the technology. It rarely is. It's usually a person, a process, or a moment of access nobody accounted for.
If you hold any crypto, this is the story to actually read all the way through.
That's the week. If you want this kind of AI rundown landing in your inbox every week, make sure you're subscribed, and follow me on LinkedIn, Instagram, and TikTok so you don't miss what's coming next.
Hit reply and tell me: which of these five surprised you the most? I read every single response.
Jenny
P.S. If you want a weekly version of this specifically focused on AI, my other newsletter, The Women AI Weekly, is built exactly for that.