AI's Economics Don't Make Sense
Summary
Ed Zitron's piece argues that the economics of generative AI are fundamentally flawed, claiming that subscription models conceal the true token-based costs of running LLMs. It examines GitHub Copilot pricing, OpenAI/Anthropic burn rates, and Stargate Abilene to argue that AI infrastructure is subsidized and debt-fueled, not intrinsically profitable. The article calls for price transparency and questions the long-term profitability and business viability of AI platforms.