AI data centers are currently our most viable lifeboat, he said. Hmmm.

Usual notice before getting started: If you like what you find here, thanks in advance for your subscription (possibly but not necessarily paid) or, even better, one-time donations for any amount, or consulting, writing or speaking work.These days, everybody loves (or needs) to hate data centers for AI, and with good reasons: assuming they are built in the quantities AI companies want them, which there is plenty of reasons to doubt, they will suck up really huge quantities of energy, money, water and human creativities that there are thousands better ways to use. A few days ago, however, someone called Brooks Fiesinger (of whom I know nothing except the post mentioned below) made on Facebook the opposite point, which is quite interesting as food for thought, if nothing else.Here is a synthesis of Mr. Fiesinger thesis, partly reformatted as a numbered list to make it easier follow the comments, and counter-proposals, that I make right after it.THE UNCOMFORTABLE TRUTH: We are running…