RawBooty.com | Price ≠ ΣFCFₜ/(1+r)ᵗ

Impact of AI - This *is* important but it's complicated, and thus widely misunderstood. I've been online daily since the early 1990s. The idea that "AI doesn't drive traffic" misses what's actually happening. AI is making traffic more efficient and more intentional. That has massive implications for a lot of businesses. Anyone who has been using the internet for more than a decade knows search quality deteriorated. Finding useful information became harder, often impossible. Websites became filled with SEO spam, affiliate content, ads, and paywalls. Monetization has corrupted much of the ecosystem. $RDDT skillfully filled part of that gap. In many ways, it became a modern version of Usenet, forums, and bulletin boards. If you wanted an authentic answer, you often ended up adding "reddit" to your search query. But that wasn't really a solution. It was a workaround. What AI has done is make research useful again, for the moment. (Don't worry this too will likely get corrupted and become useless.) Imagine you need a toothbrush. Twenty years ago, you might have searched Google, opened twenty tabs, read reviews, and compared products. Today you ask an AI: "Who makes a good toothbrush?" or "Where should I buy one?" Instead of visiting twenty websites, you now visit only two, and that's the key point. AI isn't eliminating traffic. It's eliminating unnecessary traffic. If you're an advertiser, that's both wonderful and terrifying. It's wonderful if you're the category leader, the lowest-cost provider, or you have a genuinely differentiated product. AI becomes a highly efficient referral engine. Importantly, you now have an opportunity to lower ad spend and get the same or better conversion. (eep!) It's terrifying if you're in the middle because the middle increasingly disappears. Consumers no longer need to sift through ten nearly identical options to find the obvious winner. Ad spend becomes highly problematic, as this is new territory to navigate. How do you gain awareness or market credibility? The irony for Reddit is that much of the training data used to create modern AI came from the very platform that was serving as the internet's research layer. In a sense, Reddit helped create the tool that may ultimately reduce its importance. My prediction is simple: AI will disproportionately benefit category leaders and unique businesses while compressing everyone else. The winners get more qualified traffic. The middle loses visibility until they learn to adapt. Historically, a significant amount of internet traffic and purchasing behavior was driven by buyer fatigue. People would spend thirty minutes researching a product, become frustrated, and eventually settle for whatever happened to be in front of them. That's how you ended up buying the $0.95 Fu-Long-Duck toothbrush whose bristles fell out in your mouth, instead of finding a genuinely superior product from a lesser-known manufacturer. Platforms like Amazon (unintentionally?) reinforced this dynamic by overwhelming buyers with thousands of products of varying quality and then largely rewarding whoever became best at manipulating the platform's search, reviews, and ranking systems. AI changes that equation. If it works as intended, consumers spend less time searching and more time selecting. The result is fewer clicks, fewer page views, but potentially better purchasing decisions. This has profound implications for companies like $GOOG $META $RDDT and most others. For some it's a boom time, for others this is going to be a death sentence. As for me? Well, I'm actually quite bullish on this change. It addresses many of the complaints I've had about the internet since the early, pre-Google era. As an investor your job is not to really think through these implications and try to avoid the losers and find the winners. Big change is coming. Whether AI eventually follows the same path as search and becomes corrupted by incentives, advertising, and optimization games remains to be seen. But for now, it is making research useful again. That's a topic for another post. hope this helps!

$GOOG Serious question: At some point, does scale stop being a moat and start becoming a liability? Search was one of the best business models ever because the marginal cost of serving another query was nearly zero. AI is different. If every interaction has a real per-token compute cost, and usage scales faster than monetization, then the economics start looking less like software and more like a public utility. Let's consider the business of selling electricity: Imagine a world where electricity is free for almost all use cases. Only a small subset of consumption is metered and billed. Demand would explode. Usage would explode. Infrastructure spending would explode. How do you ever earn an adequate return in this scenario? Now, what if so much capital must be deployed to support the infrastructure that the paying customer base simply isn't large enough (doesn't exist) to subsidize the non-paying users? Or worse, what if the revenue generated from the paying segment isn't even sufficient to cover the depreciation expense of the buildout? The assumption embedded in most AI forecasts is that usage growth and monetization growth are highly correlated. What if this assumption is wrong?