Table of Contents

1. Introduction
2. The Good
3. The Bad

Introduction

I’ve been working on my NPM library, and I was going to write about it. But then I saw this YouTube video about AI and YouTube, and I realized that I wanted to write a blog post about it. There seems to be two ideas, one that would be for producers and one for consumers, both similar and useful for AI.

1. AI-based creation of videos that could be based on previous material, topic chosen by the creator.
2. AI-based creation of videos with subjects/content chosen by the consumers.

It worries me and doesn’t at the same time. Whoever came up with the idea doesn’t really understand the human condition. They understand things through the lens of consumption and production, which, I guess, isn’t hard to understand. There’s that famous Upton Sinclair quote:

“It is difficult to get a man to understand something, when his salary depends on his not understanding it.”

People who make money do not think of problems in the same terms that normal people think of them, they think of economics, costs and income. It is one of the first things when you get into the real world as a programmer: what is the business cost, what is the business need, what is the value it brings to a business, etc. On your resumé, you’re supposed to use numbers such as: optimize the database and decrease query time by 30%. There is something that agrees with this view of programming when people want to replace programming with AI. And you know the but is coming because of how I phrased it. Here’s it is: there’s another side to programming, just like there’s another side to how humans exist.

The Good

Before I go over the negative, let’s talk about what AI can do. I think it’s a good way to learn something. I’m talking about something where you already have a decent amount of knowledge in an adjacent field and want to quickly learn something where you can fill in the details if you don’t know. Especially if that field is easily deterministic. For example, if I want to learn a new programming language or get a basic idea, hey, it’s great. Well, it is good for a programming language that is well known like C#, but not something like Zig. And yes, I did want to get help with Zig, and AI did not help me much.

One thing to keep in mind is that when you learn something new, what you aren’t looking for is factual knowledge. What you want is a human connection. When you learn a foreign language, it isn’t to be used in isolation but as part of a community. Humans write and record their experiences and knowledge in a way to preserve that knowledge, but they do so in their own voice. Over time, we have learned to make our recording and teaching as generalizable so as to teach as many as possible. And those that do not know the process confuse the end product for the reason of its existence. But academic teaching isn’t that great, just to get you to the point of learning the cusp of something. If you want to learn something well, you have to make a connection.

“Wait,” I can hear you screaming. “But what about people who get doctorates and such? What about you and your masters? Didn’t you learn from all those academic texts, articles and books?” And yes, I did. But that’s after I learned the subject matter well enough I could understand the humanity behind who wrote those things. I spent years face to face with teachers who spent time to connect with me, to induct me into their plane of reasoning.

I don’t hate AI just because it exists. I want to embrace a reality where the tedium of creation and the consideration of the “business” aspects is secondary to humanity. I frequently use it, in fact. I use GitHub Copilot all the time and ChatGPT for questions I have. For example, Copilot wrote some unit tests for my package which were rudimentary and didn’t fit well. And today I tried to use ChatGPT to figure out how to get a custom subdomain for my package on GitHub Pages, which wasn’t great. But don’t let me discourage you, there have been times that it has been helpful. The autocomplete often cuts seconds off how much I would have to write, often minutes across a whole day of work.

I was talking to my brother about its uses, and it’s becoming evident that how useful it is to you is inversely proportional to how familiar you are to what you are doing. I have gotten good at anticipating my needs and designing larger APIs and systems in my head, and AI isn’t at the point of doing that type of work. But if I need to figure out how to do basic layout stuff and make another corporate-looking website, well, it can do a lot of things quickly. It can provide a lot of business value for little cost because business does not care about having a unique design or something unique. Unless it provides business value.

I can feel the rebuttals. The nice things I have wouldn’t exist without people having to pay attention to business, to working on real, practical things. They wouldn’t exist without businesses selling them to me. After all, I have said that art in a medium is only interesting because of the limitations the medium put on it. But here’s my response: nah. Do I believe that some people are motivated by profit? Yes. I know they exist, and I’ve known them. But the vast majority of people who are engineers, teachers, writers, etc. do it out of passion or the innate drive to create. They bow to other concerns because we’ve decided that you must be creating value for someone in order to live a stable life.

In fact, the people who pursue a field just because of the salaries are often ridiculed. In programing, everyone knows they will not create anything good and will at best be drones in a hive because they don’t have the spark that will let them come up with a unique solution to a unique problem. That part has nothing to do with profit. And that’s where AI is: it is great at being a drone. Sometimes we need that, but often we don’t. The part that AI is good at (so far) is the trivial part. We should use it to reduce tedium, but those who seek to use AI as the ultimate solution do not know how to ask good questions.

All the tools to create this website and blog you’re reading? They’re free tools! But Gatsby is based on React, which is sponsored by Meta. Here’s the question: would the people who created React not have created it had Meta not paid for it? Certainly the same product wouldn’t exist, but Vue, Ember and Svelte exist without major corporations running the creation of the frameworks. The money Meta gives the team solves a problem that our economic system created in the first place: creativity cannot exist in its own right because people need money to exist.

The Bad

I didn’t write this just to crap on AI. After all, it is a tool in a… toolbox. One day it may be able to wield the entire toolbox, but I don’t think that’s going to happen with the current technology. And if it does, it will be a one-off solution that requires so much computing power to compile that it will take years and be an incredible drain on resources. And like all things of such a grand scale, like the Manhattan project, it should be undertaken by the government as a public tool, not as the tool of private enterprise in seek of profit.

Why I wrote this article is to think about the philosophy of our use of AI. In my life, I keep going back to when we studied Walter Benjamin in college. I had never taken philosophy, but that one class I took, the one I had to take in order to graduate with honors (summa cum laude by the way), was one of the best I took in my entire time in college. Benjamin wrote about the mass reproduction of art in society. It was an era where you were going from a stage play to a film of Hamlet. You no longer had to care about the individual actors in the individual production, how they transformed the material from words and concepts to reality. Or how an actor needs to only create an individual shot instead of acting out the whole scene and then whole act. The whole concept of doing things in post is anathema to the individual touch that an actor adds. As an aside, what would be his take on ADR?

Once you created a means to mass reproduce the creation, it no longer had the individual touch of humanity it once had. After all, when you look at a jpeg of The Birth of Venus, you’re no longer considering the hard work and toil the artist put into the canvas. And if you’re reading an ebook rather than a medieval manuscript inked by monks, you’re no longer considering the scribe who drew some ridiculous design in the corner because they were bored.

If you couldn’t tell, that last example was a bit over the top for a reason. As a society, we have accepted mass production, and there are advantages of it. It exists to increase accessibility: the ability to gain knowledge that was hard to find previously, the ability to experience something that we couldn’t be there for, etc. And in some ways, AI is supposed to do the same thing, to increase the accessibility of greater knowledge. But that’s not how it’s being used. It’s used to imitate a creative effort without any of the soul. It’s the champion of mediocrity without soul.

It all comes down to this fact: those that are putting AI in these positions of creating things do not understand humanity. They are not humanists, and they do not see beyond their nose. They do not see the value of the blossom of human creativity. They do not understand the value of community that AI generated videos will destroy. I question the people who do not think for a second about the moment of human invention. We decided several hundred years ago that creativity was the most important matter of all in our human existence. It is how the age of invention began. We should never, ever forget that business and money works in pursuit of human creation, not the other way around. And if we do, we might as well leave our cities and our inventions behind: the human race will have achieved nothing.