Kids These Days

October 1, 2026

A good friend of mine says to me "Have you considered how much water you're using?" as I talk with an AI to improve my Spanish skills.

I hear about datacenters taking electricity from farmers as I create flashcards and podcasts about classes I'm taking and subjects I'm learning.

Yes. I have considered these things.

There are many potential negative externalities to AI.¹ Yet, I do not believe that my decision not to engage with AI tools will make a meaningful difference in those externalities, especially when these companies have somewhat fixed amounts of compute. Granted, there may be issues with this way of thinking (for instance, no individual votes matter?), but on the whole I believe that AI, when used properly, can be hugely beneficial towards an individual.

Given the apparent ethical case to not use AI, many otherwise ethical people may not see these benefits, causing them to potentially lose future decision-making power to those who do.

Many of the luddites, much as one may read about them in a history textbook, ended up in penal colonies in Australia, where they certainly did not have much of an impact on the consequences of the Industrial Revolution.

A twisted version of this shows up in schools. Students that violate rules to use AI to assist with busy work that does not specifically aid their learning may actually see more success in a course than those that do not. Note that I do not say "do their homework for them" or "cheat on a test." However, blanket bans cover both types of behavior, and can force otherwise well-meaning students who would never cheat to choose between rule violation and a simply more effective route to knowledge.

All this to say, I do not believe that the vast majority of opinion-holders on AI appreciate the complexity. I only know that I certainly do not.