Much discussion of AI's future extrapolates from current approaches, but things like pre-training and RLHF can become archaic surprisingly fast. One day there will be no "pre" and no "prompt," and enduring weight-level updates will happen for as long as "life" continues, just like with humans. So you have to understand, the idea of "values" that are baked into such a continually plastic system means something different. Try to extrapolate too far from what we have today, and you risk being deceived.
By the way, such a genuinely plastic system would be quite intriguing. "Context length" would no longer be a thing and we would have for the first time in effect a kind of "immortal learner" that can keep on updating itself forever and that never dies. We're clearly not there today but there is a vast space of systems yet to be imagined. When these new paradigms are realized, much of what we think about today will be moot, regarding safety or otherwise.
@austinramzy It's not only about foolings. The more Zelensky has to prove his whereabouts, the likelier some mercenaries can succesfully assassinate him
@hardmaru@yangyang_cheng This is classic authoritarian defense. "You are free within some boundary". But they never discuss where the boundary should be appropriately set with the people.
It is important to realize that the brain is an active learning machine. It doesn't need constant external inputs to learn. Instead, it can learn in spite of lacking external supervision. Consolidation of memory and knowledge during sleeping is one example.
Recalling and explaining the popular internet phenomenon of "The Dress" with the help of some graphical tools https://t.co/lF7BunC77b [source of the gif: https://t.co/SkYkQDwYqH]
"Welcome to the mental prison that most Chinese live in for decades." -- The world, stop playing along with this madness. We, as a human race, have fought so hard to preserve and advance freedom (including a world war). Let's not slip back. Excellent work by @tony_zy and team!