You listed a few interesting things and seemed miss an important one.
I might be a bit wrong in what I’m about to say, but the basics are right. Meta released a chatGPT like LLMs source code and had their weights leaked. Their model is named LLaMA.
People have been messing around with LLaMA inspired LLMs on their personal computers thanks to meta for months now.
Bad actor LLM bots are now a hobbyist level task. The fediverse is showing signs of not significantly caring or trying. Imo, Lemmy instances aren’t ready for this.
You listed a few interesting things and seemed miss an important one.
I might be a bit wrong in what I’m about to say, but the basics are right. Meta released a chatGPT like LLMs source code and had their weights leaked. Their model is named LLaMA.
People have been messing around with LLaMA inspired LLMs on their personal computers thanks to meta for months now.
Bad actor LLM bots are now a hobbyist level task. The fediverse is showing signs of not significantly caring or trying. Imo, Lemmy instances aren’t ready for this.
https://ai.meta.com/blog/large-language-model-llama-meta-ai/