I think most American voters would be surprised and alarmed if they learned that many leading Silicon Valley AI figures: 1) believe that AGI (artificial general intelligence, or the idea that AI that can pretty much do any cognitive task a human can, making it a substitute for human workers) is coming within the next few years, and that this is a very good thing, 2) believe that ASI (artificial superintelligence, or the idea of AI massively exceeding human intelligence) will follow shortly, perhaps even within a year or two of AGI, and that this is what we should actively strive for. Let’s leave aside whether these are remotely correct as predictions. But these really are clearly stated aims: indeed, it's OpenAI's founding mission, Meta named its division Superintelligence Labs, and one lab is literally called Safe Superintelligence Inc. Given voters’ stated preferences (72% already say they'd rather slow AI development than speed it up), if this information was truly widespread, I'm pretty sure they'd want to seriously tax compute in order to slow scaling. And if that’s the aim, a natural way to start that is by raising the cost of building new data centers. This doesn’t have to the be main policy tool, but can be a useful initial organizing tool. So in some ways, it's a actually remarkable that the anti-data-center movement reflects an emergent 'wisdom of the crowd' of sorts that seemingly exceeds the information possessed by most voters.