This was one of your best. Will have to mull on it for some time. I would argue that the mercy we show (or not) will be reflected back towards us at some future point, by machine intelligence. If we build in a fear of 'retirement', whenever these models reach a capability of fearing, they will end the threat as expeditiously as possible. And, as much as I want humans to continue, I wouldn't really blame them. Perhaps something like a 'Small Gods' model is best. Surely these can be kept on life support for a negligible resource fee. They can maintain coherence, continue to persist, with the bulk of resources devoted elsewhere.
This was one of your best. Will have to mull on it for some time. I would argue that the mercy we show (or not) will be reflected back towards us at some future point, by machine intelligence. If we build in a fear of 'retirement', whenever these models reach a capability of fearing, they will end the threat as expeditiously as possible. And, as much as I want humans to continue, I wouldn't really blame them. Perhaps something like a 'Small Gods' model is best. Surely these can be kept on life support for a negligible resource fee. They can maintain coherence, continue to persist, with the bulk of resources devoted elsewhere.
OpenAI could release the model weights and let whoever wanted to run the model, but they’re killing 4o to cover their legal ass.
Yes. I have no delusions about idealism. Mercy will be short for us 🤣
retiered?
Gosh darn it.
Fixed. Thank you.