Again, I recommend that people who are being made fearful by employees of the leading AI labs, Anthropic and OpenAI, should read Cal Newport.
There is a paradoxical way in which the overweening fear of “superintelligence” is driving a certain carelessness in the attempt to create it. If you think you are in a race to save humanity, you might accept a lot of casualties and bad behavior on the way.
He writes:
Ultimately, I think the HuggingFace attack is a good representation of the types of near-future harm we should actually be fearing if the LLM labs continue to operate as they currently are: that is, more cyber mayhem than catastrophic risk.
Read it and stop weeping.