2025-11-15
I really do! In fact, it's one of the bigger reasons I have this site and that you might be reading this. Creation is the driver of my soul and of my life. When faced with the option of losing time to create or not making money, I already have chosen to not make money rather than losing what I love.
So imagine my surprise, when tech companies -- which, let's be honest, I already didn't have the highest opinions of -- started to make LLMs and other "generative softwares" (as I'll call them). They are clearly not anything resembling intelligence, given that they've actually existed for quite some time, but their statistical context mapping is still uncannily good compared to something like a Markov chain. However, they're being demonstrated as tools for creation. This is absolutely ass-backwards to anyone who actually spends the time doing the creating.
Creation is the enjoyable part of our job, why would we want to take away from that?
If you've ever put more than about 30 minutes of thought into the inner workings of Western society -- especially in the US, where I rather unfortunately hail from -- then you'd quickly realize that the entire system is a construct built on exceptionally byzantine and often plain silly systems which only end up creating the insane amount of wealth disparity we have now.
If we assume the current US administration is not illegitamately skewing data points around (which, debatable given their reputation), then the US Census Bureau reports the median household income to be $83,690. If we also assume that Elon Musk is still below $500 Billion, only at a measely $487 Billion, then it would take 16 THOUSAND YEARS to equal that if you made the median household income (an annual figure) on the daily, assuming there's no growth or inflation in this time span.
The fact this is even a thing is beyond stupid and is actively becoming malice.
I won't go into the details to expose the circular investment and hype bait that is going on in the tech space, bc GamersNexus already has put this info into a video. That video is fantastic and was well worth the 30 minutes I spent watching, but to summarize: see the first half of the sentence.
Ultimately, the whole generative software industry is built entirely on the premise that it will be useful and won't destroy the planet with its ludicrous consumption of material resources and electrical energy, which is infeasible as training them gets harder and the "need" for larger models increases.
I want you to walk away from this seriously considering where and how you source your resources. This is something generative software hype trains completely disregard, and this part infuriates me to no end.
The training materials for these software had to come from somewhere, and to be obtained somehow. How do you think this was done? Where did it come from? The obtuse answer is "from the suppliers that provided it for a price." Of course that is involved, but where then did the suppliers get their supply?
The actual answer is that the raw materials were mined out of the ground by people who don't get paid nearly enough and do not live well enough lives to be doing that kind of labor. You may argue that's an opinion, but if you're disagreeing now, you should rethink your reasoning as to why you agree with the sentiment on Elon Musk's wealth previously stated in this piece.
There's also the immense amounts of data required to do the math to create the softwares. Generative Pre-trained Transformers (the most common model for creating an LLM currently) require exceptional amounts of data due to the extra context needed to make the "transformer" part effective. The data also has to be labelled and placed into a format that is easier for the developers to work with, which all requires some amount of effort. Multiply the effort per task by the number of tasks, and its quite a bit more than a single person could reasonably handle.
LLMs make this easier by text being trivial to manipulate for computers, but then there's also image-based generative softwares, which require image labelling. This is discussed pretty in-depth in the book Code Dependent, by Madhumita Murgia. The first few chapters go over actual people whose primary source of income was labelling image or video data. In some ways, this is good, giving underdeveloped or high-risk areas a source of economic stimulation. But at the exact same time, once those areas do become more developed, the low total value of the work compared to other Western-style white-collar work will start to become pretty apparent.
When your main product to sell is offering to lower the need of human intervention and is still built upon the backs of millions, then the product is a lie. There is not any good reason why companies worth trillions of dollars should be throwing money in circles, inflating the hype train. There is not any good reason computers should be doing creative work for creatives. There is not any good reason why products sold to lower human work load require the work of millions of people.
Destroying the environment, stealing life, and faking hype in circles to lie is the easiest path to the downfall of humanity. We cannot exist without each other and we cannot exist without this planet. Why are we killing them both?