With automated traffic now dominant and predicted to be 1000x human traffic, the ad-based internet model is broken. Bots don't click ads, so Matthew Prince argues a new model, likely micropayments for data access, must emerge to fund the required infrastructure for this new reality.
AI companies want to 'fill the holes' in their knowledge, making unique local news (like a new restaurant) or niche community data (like Reddit) more valuable for licensing than another story on a national event. This creates a new economic incentive for local journalism and unique content over high-volume, substitutable news.
Cloudflare will set the default for its free customers to block Google's AI training crawlers. By controlling over 20% of the web, this action creates a significant hole in Google's data, applying technological pressure to compel Google to negotiate fair payment terms for the content it uses to train its models.
Cloudflare categorized staff as builders (engineers), sellers (sales), or measurers (audit, finance, middle management). Layoffs primarily targeted "measurers" whose roles in auditing and reporting could be performed more efficiently and objectively by AI, while the company continued to hire builders and sellers.
Cloudflare observed that the most junior employees (digital natives) and most senior employees (confident enough to bet on paradigm shifts) rapidly adopted AI tools. Mid-career staff, trained in established processes and rules for success, tended to resist the change, creating a "U-shaped" adoption curve within the organization.
While humans share correlated biases from working in the same culture, AI models have different, uncorrelated biases. Cloudflare uses this to its advantage. By analyzing performance data, AI can surface high-performing junior employees who might otherwise be overlooked by the conventional management chain, enabling better talent recognition.
By providing tools to block AI crawlers, Cloudflare creates a constrained supply of previously free data. This manufactured scarcity is the foundation for a new market where content creators can charge bots for access, shifting Cloudflare from a security provider to a market-maker for digital information.
By using AI tools to help managers surface issues and track team performance, Cloudflare found it could increase the average number of direct reports per manager from ~6 to ~12. This reduces the need for middle management layers, directly leading to a flatter, faster, and more nimble organizational structure.
Drawing parallels with musicians who often value fame over fortune, Matthew Prince argues that solving the financial payment model for content isn't enough. The next great challenge is creating systems of recognition, as creators will stop contributing (like on Wikipedia) if their work is only consumed anonymously by AI.
Matthew Prince argues that waiting for industry consensus before making AI-driven layoffs is cowardly. He believes it is kinder to employees to restructure early, giving them a better chance to find new roles before the market is flooded, even if it means the leader must face public backlash and personal threats as a first mover.
Similar to how Spotify pools subscription fees and pays artists based on streams, a new model could pool fees from AI companies for data access. This money would then be distributed to content creators, creating a sustainable ecosystem for information that is currently being scraped for free and without compensation.