LLMs ("large language models") need a lot of input to generate their statistical text models. Orders of magnitude more than is contained in a single webcomic, even one as large as Achewood. The way this works is, roughly, you train a "base model" on approximately the entire internet. Then you take this "base model" and you add additional training on top to give it a specific flavor, such as "Ray Smuckles". The article alludes to this:
The first challenge was formatting the data in a way the language model could use. Then, there was the matter of picking an underlying language model to train with Rayâs voice. OpenAIâs ChatGPT was a little too sanctimonious for Ray, who likes to color outside of the lines, Hall says. They wound up using a fine-tuned version of OpenAIâs Davinci, which Hall estimates is about 60 times more expensive than ChatGPT.
So, this is not just a matter of "he's only using his own writing so it's fine". The model Onstad is working with is exactly as plagiaristic as anything else OpenAI has put out, it just adds a layer of Smucklesiness on top of that. Whether you think "training a statistical model on the entire internet without authors' consent" is specifically plagiarism, otherwise exploitative, or totally fine is up to you. But you can't draw a clean line and say "training AI nonconsensually is bad but what Onstad is doing is okay."
This is one reason why I think this technology is so harmful and the discourse around it is so cognitohazardous. It's poorly-understood in ways that make all sorts of category errors and ethical invisibilities possible. My view is that total rejection is the only acceptable way to behave towards it, because the well is just entirely poisoned by the tech industry trying to both normalize and mislead people about it.
If an LLM was involved in making new Achewood then that's Zombie Achewood. It's dead to me, and so is its author. I won't support it or promote it or feel any excitement that the webcomic is back. Because frankly my position on this is not a nitpicky evaluation of the ethics of any individual case, it's that I want this bathwater out of my house and I don't give a shit how many babies go with it. We can't even begin to have a discussion about the 'best-case scenario' until, frankly, every tech CEO is in the ground.
