Hope that open source LLMs take off more than closed source proprietary ones.

data in lemmy is fundamentally open for scraping. Scraping itself isn’t fundamentally bad, but it depends where we feed our data to. I am at ease with my data being used to train an open source LLM (open dataset, open algorithm, open training procedure, etc) rather than used to train microsoft Bing’s new chatbot. At lease we need an open source LLM to be able to compete with the GPTs.