How does the language of contempt shape the public’s understanding of ppl and issues they, understandably, know little about?

It starts w polemicists like this site and works its way into receptive minds at big media who are more interested in adopting political narratives than in covering the news

Translation AI is much smaller and more efficient, though. I'd put it in the class of SLMs: Small Language Models.

Translation is powered by the same tech as LLMs, but with a different architecture and a much smaller scale. Both are Neural network based language models, but LLMs are trained to predict the next word while Translation AI is trained to transform sentences into different languages.

The one place I've found people who actually understood and cared was a womens' CSA survivors support group: the kind TERFs are virulently opposed to us being allowed in. I was the first Trans women to join, and it helped me so much.

I’m a disabled homeless trans woman living in a motel in the short term. If anyone can help me out with funds to extend my stay and cover necessities, I would be immensely grateful. My links are in my bio. 💕💸💸💕

#crowdfunding #disability #homeless #transcrowdfund #mutualaid

That was where the field was heading, until billionaires injected their unneeded control and capital to create OpenAI.

Even LM generalized pre-training can done efficiently in reasonable sizes on a handful of consumer GPUs using public domain datasets, and then fine tuned to be excellent at specific tasks. Hugging face is filled with open source examples with a variety of sizes and specialties.

My primary AI work lately has been focused on document processing and sensor processing algorithms, which have virtually nothing to do with what is generating all the AI hate except for the underlying technology.

I minored in it as an undergrad in the late '00s, used them in computational neuroscience work in the early '10s, and finally transitioned into consulting around 2018. They are fantastically useful and efficient as tailored tools for specific purposes, rather than as massive general purpose ones.

They were useful for so many things other than replacing human beings, and without needing massive data centers to power them. They still are.

I was making a living as a neural network machine learning specialist for years prior to billionaires realizing that if they monopolize the world's GPUs and funnel the electricity consumed by a small city to them, they might be able to take even more power from the masses.