It is hard, especially on the Internet, to admit when you are wrong.

AI is a contentious topic, and I have many thoughts on it, depending on the field, use case, and phase of the moon. I'm (sometimes) happy to talk about it in person, when I can be sure that I'm having a conversation with a real person.

Since its arrival on the "tech scene" a few short years ago, it's dominated countless conversations I've had with friends, family, and strangers alike. Having some kind of opinion - any opinion - is table stakes.

I started from a policy of not thinking about it, and have gradually proceeded from there. By now, the opinion that I have arrived at is that AI is a fast-rising tide. I have spent more than a decade writing code, and there are people barely 6 months past running their first program who can far outdo my best production efforts.

I am proud of what I am capable of on my own merits. I think that I have written code that can be considered beautiful. ^[1] I want to be a person who puts beauty into the world, whether through code, theorems, photography, or just appreciation of a nice view. At the same time, I want to be pragmatic, or at least reasonably so.

I can no longer ignore the efficacy of AI when it comes to programming. It is simply too powerful, and too prolific, to be cast aside. This is not to say that AI will kill us all, or that there will never be a place in tech for humans. Rather, I have come to accept that AI, when used as a tool, is tremendously useful. For a long time I refused to use it under any circumstances, and I don't think that will be the case anymore.

Starting today, I'm going to cautiously use AI. I've written a NixOS module to enable it on my system, but gate it with a lecture ensuring that the user (me) accepts the responsibility of using a very powerful tool. Like guns, cars, and airplanes, AI is a machine representing the culmination of thousands - hundreds of thousands - of person-hours of work, and that makes it very powerful. It also makes it dangerous.

I do not want to - and will not, to the best of my ability - use AI in any project that I am pursuing for my own learning. I know and understand that doing so is detrimental both to me and to my work. However, in circumstances where it makes sense, I will use it to accelerate my work. I understand that doing so feels like a betrayal of beliefs and principles I have espoused in the past. I am using this opportunity to (very minimally) publicly admit that I have changed my viewpoint.

I am, as always, choosing the life I choose to lead. I will think before I act and after, and I will continue to improve my understanding.

<3ula