An interesting outside view I stumbled recently concludes:
I think the typical pattern is for human artifacts to be 2-3 OOM less efficient than their biological analogs, measured by “How much energy/mass is needed to achieve a given level of performance?” Manufacturing costs differ by a further 2-4 OOM.
An interesting outside view I stumbled recently concludes:
Thanks Misha. Can you share who the author of that document is? Looks like it might be Paul Christiano but I want to be sure.
I think he is the author.