Link to the article: https://www.qualcomm.com/news/releases/2023/07/qualcomm-work...
Can someone explain to me why Meta's approach is responsible? I mean, I applaud Meta for "open sourcing" the models, but don't they contain potentially harmful data which can be accessed without some kind of filter? Let's say retrieve instructions on how to efficiently overthrow a government?
Your license to use Llama can be revoked if Meta investigates and deems your action to be against the code of conduct[1]. I suppose if you continue to use the model without a license, you may be open to subsequent legal action in uncharted case-law territory.
1. https://github.com/facebookresearch/llama/blob/main/CODE_OF_...
lol the examples people give for AI safety are always so ridiculous. Here I am thinking I’m bad at estimating when I misjudge a week’s worth of work for a month and this guy thinks he can overthrow a government by following a few thousand word plan written by an LLM.