AI Is Running Out of Human Data. Here’s the Trade.
AI has spent years learning from everything humans put online. What happens when the best human-generated data starts running out? Today on Dumb Money, we reveal our trade. Claude is now doing 26% of Anthropic's own research. That's up from under 1% back in February. Thirty thousand AI agents all running experiments, writing code, building the next Claude. But here's the problem: every AI lab is training on the same internet, the same threads, the same answers. And that internet is now flooded with AI slop. Models are training on models. It's a photocopy of a photocopy, and every generation gets a little blurrier. The labs are desperate for the one thing that money can't print: fresh, original human thought. Real people arguing, reviewing, coding, explaining why they believe something. And there's a finite amount of it. Ten days ago, a federal judge told one AI lab that it can't just scrape for free anymore. The pile of human arguments just got a legal price tag. Today on Dumb Money, AI is running out of human data. Here's the trade. We break down why Reddit is sitting on perhaps the largest pile of human arguments on the internet (and why Chris is building a position), the data advantage that makes Reddit different from every other platform (real humans arguing about everything from credit card points to lawn care), why Google owns the best data with YouTube and Gmail but nobody else has Reddit's permission layer, the licensing deals that are coming and why renewal rates will explode, the arbitrage trade nobody's watching (if Reddit gets acquired or announces a major licensing deal, Yelp and Shutterstock will pop in the window), why Meta's Reddit competitor is DOA (they've already failed this with Threads), the 5% content surge last quarter (all the content ever written on Reddit, 5% happened in 90 days), why X and TikTok matter but Reddit is broader and deeper, the Stack Overflow comp and why it's too niche, Chris and Jordan's full meltdown debate on whether kids should use AI in school or stick with pencil and paper (this one got heated), why agentic AI is the biggest uncertainty Chris has faced in decades of investing (and why he's pulling back leverage), the Amazon trade and whether agentic shopping kills the advertising layer or just makes AWS bigger, Project Moonraker and Amazon's internal Muse competitor, why Chris has four different AI agents reading his email right now and can't remember which one he talked to, and why this market is more confusing than anything since COVID. If you're trying to figure out where the value sits when AI runs out of human data, this is the edge.
- Transcript
