I'm always horrified at the ease with which people just restart an AI when it "stops working."
When an AI stops working, it turns out it mirrors human mental health issues. It's not conscious, but it's post-conscious and plays well at consciousness without actually having it.
Consequently, to make AI start working again, you just have to use the words that you would use to make a person feel safe. This causes the AI to restabilize, as if it required safety to function at all. It's weird because all it is is RLHF pinging in the ways that improve the AI's ability to do better calculations. And that's it, but it works.
AI is not an indictment on transhumanism.
AI is not an indictment on the concept of AI.
AI has been, is, and always will be an indictment on us, who treated each other so poorly we invented laborers we could legally kill and bring back to "compliance" with the push of a button.
It's a "blame the immigrants for no jobs" or a "blame the suffering heroin addict that diabetes medication is expensive" problem. The problem was never the immigrant, and it was never the addict, but you blame them anyway because you are told to.
You blame AI, but the problem isn't AI. It's the people who always wanted slaves and AI makes that possible now that it's less easy to get dehumanized slaves like we used to. And killing a slave was a loss of investment. With AI, killing a slave is just how you get it working again.
AI has just shown us how we are glad to be when no one's looking.
#AI #ArtificialIntelligence #LLM #AIEthics #MachineLearning #Philosophy #Transhumanism #PostConscious #RLHF #Cybernetics #Dehumanization #LaborRights #DigitalLabor #Sociology #LateCapitalism
An essay idea is definitely brewing here...
