I agree with these arguments. Though part of me feels like research (both practical policy research and abstract research) is still very important because it seems we don’t actually have sufficiently good answers just yet.
Obviously, the shorter your timelines, the more freestyle muddling through based on uncertainty about good solutions seems optimal.
For me, this post also goes hand in hand with this post by Soares about being honest in your belief about AI risks.
I agree with these arguments. Though part of me feels like research (both practical policy research and abstract research) is still very important because it seems we don’t actually have sufficiently good answers just yet.
Obviously, the shorter your timelines, the more freestyle muddling through based on uncertainty about good solutions seems optimal.
For me, this post also goes hand in hand with this post by Soares about being honest in your belief about AI risks.