A few interesting ideas from Venkat Rao's Beyond Szabo Scaling:

"I propose societal expressivity as the right quantity to try and maximize. Loosely, for every problem at every level, build the most capable global computer (human + machines) we reasonably can, leaving a lot of surplus expressivity and power to work with. What's more: This is in fact what we've actually been doing for 200 years, overbuilding societal "computers."

"Douglas Hofstadter, for example, offered the dismal idea that "apathy at the scale of individuals is insanity at the scale of civilizations," an epigram that is pessimistic about the quality of collective cognition and care at scale rather than trust, which makes it an epigram that we must skeptically reconsider in light of AI advances and its potential for addressing insanity at scale (so far, we've only been clutching pearls about how AI causes insanity in lonely, atomized individuals)."

"Libertarians actually prefer this of course. They prefer direct human sociality to remain small-scale, enduring and intimate, leaving social scaling beyond the Dunbar limit to more indirect and impersonal mechanisms ranging from public-key cryptography to markets to voting. Rather paradoxically, as we've come to realize in the last decade, they are also actively eager to find and enthrone putative "Great Man" types in unique positions as scalability hacks. Libertarianism in actual practice appears to be a combination of trust-minimization and demigod-construction."^[ad]