I believe that this would be the optimal video to create for the optimization target “reduce probability of AI-Doom”. It seems (barely) plausible that someone really smart could watch the video, make a connection to some obscure subject none of us know about, and then produce a really impactful contribution to solving AI Alignment.
I would be excited to see Rational Animations try to cover the Hard Problem of Corrigibility: https://arbital.com/p/hard_corrigibility/
I believe that this would be the optimal video to create for the optimization target “reduce probability of AI-Doom”. It seems (barely) plausible that someone really smart could watch the video, make a connection to some obscure subject none of us know about, and then produce a really impactful contribution to solving AI Alignment.