Techno-Fascism Led us Here: johnnydegrowth.micro.blog
@johnnydegrowth I’m not sure what techno-fascism is. I understand fascism as a populist form of nationalism or racism and I’m not clear how using AI necessarily contributes to it. But I won’t pursuie that question any more because it’s a red herring imho.
More importantly, I feel that the enemy in this post is mis-identified. The enemy is not AI – it’s the owners of the overbuilt AI systems and the purposes to which they would put the tech.
Luddites were not against looms; looms are great. Luddites were against firing skilled craftspeople and replacing them with children and Napoleanic war prisoners to save money. A loom in the hand of an artisan is a beautiful tool.
Similarly, AI does not have to use LLMs and waste so much energy. The current owners, the real enemy, are building it this way on purpose.
I believe that unless people who care more about human uses of technology than corporate advantages of using it actually get involved, we are ceding the field to oppressors.
Any tool can be an instrument of liberation or oppression. We need to nail the oppressor, not the tools.
@tbn32 Show me a LLM that wasn’t trained off of creative theft and we can talk! I see them no different than an attempt to enclose the creativity of humans within the bounds of economics.
I theoretically agree that any tool can be an instrument of liberation or oppression, but the usage of tools in cultures and societies have ramifications, and they are designed within the context of society.
Computers, the internet, XR, and “AI” are all the tools that one can argue could aid in liberation or oppression, but I believe computers and the internet have made the world less liberated while in the hands of large corporations and I don’t see how XR or “AI” will be any different.
As far as I know there’s no “AI” that doesn’t have huge power and water draws, isn’t built off of creative theft, and that is in the hands of an entity to benefit the public good.
Practically, until a tool can be shown to be good overall, it’s hard for me to give any technology the benefit of the doubt.
Theoretically though, I agree with you.
I am not defending LLMs -- LLMs are built Large and costly as a barrier to entry -- SLLs can be employed just as effectively for most tasks -- furthermore, Deep Seek, while in the LLM class, operates differently than the established LLMs and reduces computational costs by 95% per token. We are discussing an era which should be over