I’m starting to think that we need to see AI research in the same way we see biological weapon research - a visit from a SEAL team or a cruise missile for any identified laboratory. Smash the disks, burn all the print outs!
Okay, this is hyperbolic and unrealistic, but I agree with this lion-maned YouTuber - we are really not ready.
AI as a tech is game changing, but it practically demands at least UBI (and probably some form of socialism) as a prerequisite. We, meanwhile, are still electing conservative governments! The same arseholes that will label the legions of unemployed artists, actors, musicians, coders, admin assistants etc etc as lazy and cut their benefits.
Does anyone truly believe that a tech that can replace half of human jobs is going to create happy outcomes in today’s society? Or will it just make tech-bros and scammers richer, and virtually everyone else poorer?
I haven’t clicked on the video but since you’ve said “Lion-maned YouTuber” I’m going to guess it’s Kyle Hill.
AI or no AI, the solution needs to be social restructuring. People underestimate the amount society can actively change, because the current system is a self sustaining set of bubbles that have naturally grown resilient to perturbations.
The few people who actually care to solve the world’s problems are figuring out how our current systems inevitably fail, and how to avoid these outcomes.
However, the best bet for restructuring would be a distributed intelligent agent system. I could get into recent papers on confirmation bias, and the confabulatory nature of thought, on the personal level, group level, and society level.
Turns out we are too good at going with the flow, even when the structure we are standing on is built over highly entrenched vestigial confabulations that no longer help.
Words, concepts, and meanings change heavily depending on the model interpreting them. The more divergent, the more difficulty in bridging this communication gap.
a distributed intelligent system could not only enable a complete social restructuring with autonomy and altruism both guaranteed, but with an overarching connection between the different models at every scale, capable of properly interpreting the different views, and conveying them more accurately than we could have ever managed with model projection and the empathy barrier.
