Episode #56: The Internet’s Business Model Is Cracking..What Comes Next?

00:00 Stewart Alsop introduces Cloudflare and Stratechery, with Stewart Alsop II framing the idea of network administrators versus database administrators.
05:00 Discussion turns to Cloudflare’s distributed network, AI crawlers paying to scrape, and parallels with Apple’s App Store tolls.
10:00 Cloudflare’s core functions are outlined: CDN, DDoS protection, web application firewall, DNS, zero trust security, SSL/TLS, and load balancing.
15:00 The focus shifts to Perplexity, AI scraping practices, lawsuits against OpenAI, and Anthropic’s $3,000 per book settlement.
20:00 Google’s monopoly case, PageRank, and whether AI chat is true competition for search come into question.
25:00 They recall the 1990s internet commercialization, ARPANET roots, TCP/IP, and Al Gore’s role in the “information superhighway.”
30:00 The talk explores niche magazines, ads as content, early internet communities, and conferences as proto-networks.
35:00 Spam is compared to door-to-door sales and Tupperware parties, showing how unwanted commercial attention evolves.
40:00 Science fiction predictions like Dick Tracy’s watch, real-time translation, and the future of the internet’s business model.
45:00 The episode closes with reflections on space exploration, SpaceX, Starship, and how the internet may face its own existential shift.

Key Insights

1. A central theme of the conversation is Cloudflare’s positioning as the “network administrator” of the internet, contrasting with the role of “database administrators.” Stewart Alsop II highlights how this mindset—focusing on distributed connectivity rather than centralized data—has shaped Cloudflare’s growth into a foundational layer of the web, handling massive portions of global traffic and AI queries.
2. Cloudflare’s business model is rooted in offering free protection and performance services, including CDN, DDoS mitigation, DNS, web application firewalls, and zero trust security. Over time, this freemium model has scaled into large enterprise contracts, echoing how companies like GitHub monetize advanced features while keeping entry-level services widely accessible.
3. The discussion emphasizes Cloudflare’s novel approach to AI crawlers: charging for access to content instead of allowing free scraping. This mirrors Apple’s App Store toll model, raising questions about whether such control could eventually be seen as monopolistic if Cloudflare becomes the default gateway for AI training data.
4. Broader AI tensions surface in the critique of Perplexity’s scraping methods and in the legal battles over copyrighted content. Anthropic’s billion-dollar settlement to compensate authors shows how companies are willing to spend heavily to avoid legal precedents that might restrict data access, signaling how unsettled the rules of AI training remain.
5. Google’s position is examined in light of DOJ scrutiny and the shifting competitive landscape. The conversation contrasts search’s reliance on PageRank and links with AI chat’s direct answers, suggesting that Google’s architecture is optimized for one model of information retrieval while AI points toward another, potentially disruptive future.
6. Historical parallels add depth: the commercialization of the internet in the 1990s, Al Gore’s support of the “information superhighway,” and the role of niche magazines and ads-as-content. These examples highlight how new communication technologies have always disrupted business models, with AI and the internet facing a similar inflection point now.
7. The episode closes by looking forward, drawing on science fiction’s role in shaping expectations—from Dick Tracy’s smartwatch to real-time language translation. Yet unlike sci-fi’s optimistic visions, the internet’s future feels uncertain, with questions around monetization, spam, trust, and even the existential sustainability of the current web business model.