Zitron doesn't offer much that you have to believe. His article links to numerous facts and reports. The physics of GPUs and data centers don't rely on one's be

Zitron doesn't offer much that you have to believe. His article links to numerous facts and reports. The physics of GPUs and data centers don't rely on one's beliefs. The amounts of money spent on AI versus the non-existent profits don't require belief, just looking at financial reports. The AI industry has cited "scaling laws" to justify the huge spending on AI, promising magic results just around the corner because of scaling. Zitron just shows that hasn't happened and that physics and economics will impose limits on scaling, likely long before the promised AGI or massive productivity gains from AI materialize. Newer and more sources of electricity only possibly address the environmental/climate damage from AI data centers. The problems come from power consumption and cooling, not from lack of windmills or batteries. If the thought leaders in charge of spending hundreds of billions on new data centers thought renewable energy would solve the problems they wouldn't talk about building nuclear power plants.