Agreed. If the AI capable of matching the output of 1000 office workers uses less energy than the daily lives of those 1000 office workers, then it's an efficiency win. I haven't run the numbers, but I would be absolutely astounded if people use less energy than an eventually-matured AI capable of similar output.
Long, long ago, and far, far away, I received an email apparently addressed to me, from an editor at Forbes writing a kill piece on why computers were never going to improve American business and productivity.
With some digging around, there was a senior economist of the same name nearby, and the Forbes editor had probably messed up the address, but I wasn't particularly sympathetic to his view point and at that time emails were readily discoverable, so it was clear that the editor was a computer Luddite.
What to do? I waited until the editor's stated deadline was close, and then sent a well written (I thought) response on why computers were going to improve American productivity. Even though I signed my full name, for some reason the editor never responded.
Yes, I still read Forbes; I just don't take their views as gospel.
I do think that it can be quite difficult to foresee the impact of new developments or technology accurately. Looking back and American energy transitions, wood, water, coal, steam, oil, electricity, etc., it took awhile for the benefits, and costs, to shake out. New Hampshire and much of New England got clearcut for charcoal for steel making in the late 1700s, before coal became more common. Lots of dams got built to power mills in the northeast, and then production and energy sources shifted. I think sometimes it is hard to see the forest for the trees.
In history, there have always been early adopters and the mainstream. Oftentimes with the benefit of hindsight, I think both sides often get a little wound up in consuming their own Koolaid. E.g. Tesla and Edison, AC vs DC.
All the best,
Peter
practicalpie.com
