I don't hate μGPT. I think that is an amazing piece of science! The whole idea of neural networks is amazing, and the research is cool, but LLMs are Evil. And they make us complacent. I'm angry but you get the gist (ha ha ha). Also: probability-predicted text and neural nets were lowkey disproven as effective tools, right? Gotta find some sources for that. LLMs destroyed NLP.
Recommended short reading: the paper "Refusing GenAI in Writing Studies" & more by that paper's authors.
Need to finish reading Walter Benjamin's Art in the Age of Mechanical Reproduction. It's pretty short, I should finish it soon. Will write about image-generating AI models when I get done; needless to say I think they're preeetty stupid.
take me home