AI Waste (www.kgadams.net)

A recent critique highlights the unsustainable costs associated with artificial intelligence (AI) investments, emphasizing that trillions have been spent on computational infrastructure with only minimal returns. The core technology behind many AI systems, primarily large language models, relies on intensive computations—around 10^23 floating-point operations—that are not only costly but also environmentally taxing. The discussion points to a troubling trend of "circular dealing," where companies cycle investments among each other without genuine consumer demand, creating an illusion of growth while struggling with actual profitability. The environmental implications of AI data centers are stark, as they consume tens of megawatts of power and significant water resources for cooling. While some facilities are exploring alternatives like small modular nuclear reactors for power, the challenges of nuclear waste management linger. Moreover, the constant push for advanced hardware has inflated costs for consumers and contributed to a tech supply crunch. Despite the benefits of AI, including efficiency enhancements and creative applications, the critique underscores a pressing need for a more rational investment approach in the AI sector, as the current trajectory may lead to a backlash from both the market and the environment.