“Technology doesn’t arrive directly from knowledge, but filtered through cultural and economic conditions. And the impact of that technology depends as much on the cultural perception of it as the technology itself.” – Feral Historian
The Feral Historian is one of my favorite YouTube channels. His weekly videos use science fiction as a way to examine cultural and historical issues both as an analogy for real-life history and by showing how culture is reflected through the stories we tell.
His most recent video is a bit of a detour as he reviews the book “The Social History of the Machine Gun” by John Ellis. As I was watching the video, I couldn’t help but notice a number of parallels to the adoption of A.I. that we’re currently seeing.
Many of the aspects this video details about the mechanization of war and other technology developments, and the mismatch with the social conditions of the time, line up with what we’re seeing with the implementation of A.I. systems in various fields.
Pay attention to the quotes from John Ellis’s book early in the video about the necessity to achieve high productivity to counter rising wages and the lack of skilled workers. Then consider the rising wages of the last few years and all the complaints you’ve heard about the loss of America’s manufacturing ability and the problems with our educational system.
In short, we just got a big jump in technology that’s being implemented based on immediate need, not wise forethought, and our culture and understanding of it is going to take some time to catch up. Yeah, it’s going to be a mess but it’s the kind of mess that has happened regularly throughout history. That doesn’t make it any more fun for a lot of people to live through but we can have hope that we’ll come out the other side as we have before.
My own use of A.I. continues to evolve. For the sake of my responsibility as an author and my own skills, I still maintain my commitment to human-generated content. I personally write posts like this and all those under the Rogue C# series. I write all my own code. I take complete responsibility for every word and line of code that goes under my name.
A.I. in itself is not the problem, however, and it would be silly of me to ignore the possible benefits from using it as an assistant. I do not generate code with A.I., unless I need to see an example of how a function or algorithm works or need to reformat a large amount of code and know exactly what it needs to look like but don’t want to spend all that time reformatting manually. I’ve pasted my code samples into it, asked for advice on improving the code and occasionally take its suggestions. It makes a great second set of eyes and, used correctly, a decent study assistant.
I use Google Gemini to generate the featured images for the Rogue C# series because it simply does a great job of generating the image I want with carefully worded prompts.
Is this A.I. slop? I don’t think so; I’ve seen careless uses of A.I. that I definitely call slop but this is actually an impressive product and effective accent and attention-getter for the article on adding potions to a roguelike game. It is appropriate to its use case. I still worry that some people will see this kind of promotional “art” and be turned off from the series by the tinge of A.I.. I worry more that, without an attractive attention hook, my long-form articles on programming a roguelike in C# will get a lot less traffic and have a lot less visible personality.
I’m still struggling with the use of ElevenLabs to generate voiceovers for the few YouTube videos I actually create. It makes the process a lot easier but it’s not my voice and I don’t like maintaining a clone of my voice on someone else’s server. The one video I used a voice clone for was okay but made me sound like I was a little out of it. I need to rework my video production overall and decide where it fits into my work.
This past week, I used NotebookLM to generate detailed meeting minutes from a recording that I uploaded. It might have misinterpreted one or two things so they did need to be reviewed for accuracy but no more so than those created by the average human. Again, it saved a lot of time. I recently suggested that someone get in the habit of having ChatGPT or another LLM review his correspondence for professionalism and tone because I can say, from experience, that it is useful and his writing definitely needed the second set of eyes.
There are serious issues with the way A.I. is being implemented but the use of the technology is not a reflection on the technology itself, it’s a reflection on how we develop technology faster than we develop ourselves. There are also serious issues to how we respond to new technologies. They are neither the saviors nor the bogeymen we often seek; they are tools. Tools can be used to build, they can be used to create dependency, or they can be used as weapons. It’s up to us to decide which.
