Why I don't use LLMs or Gen "AI"
a week ago
For me it's primarily the moral issues and the amount of power required. Yesterday I discovered this lovely post by benui, an experienced ui designer/programmer:
https://benui.ca/why-i-refuse-ai/ and it is a good list of why using LLMs and their ilk are bad. To be clear: I have seen machine learning used in my previous career and it is awesome. Stealing art/stories/code to make LLMs work, that I cannot accept.
I wanna write more but I'm not in a good place after hearing people say they don't care about the morality of stolen art/stories/code etc.
https://benui.ca/why-i-refuse-ai/ and it is a good list of why using LLMs and their ilk are bad. To be clear: I have seen machine learning used in my previous career and it is awesome. Stealing art/stories/code to make LLMs work, that I cannot accept.
I wanna write more but I'm not in a good place after hearing people say they don't care about the morality of stolen art/stories/code etc.
Whether you publish your writings or keep it just to yourself might be worth another debate though.
Personally I'm also not a fan of AI (LLM/diffusion) but it won't disappear and honestly I'm looking forward to the day I get the opportunity to just steamroll someone on a project because they never developed any actual skills besides prompting an AI. Most of them probably heard that "it only guesses the next most likely word" but never understood the implications of that.
But it really depresses me when actual artists start using them or some people fine-tune them to replicate a certain artist's style.