Some of us have been making that very point for the past 20 years... https://twitter.com/alexandrosm/status/1314383373163347968

To clarify: Though in favor of hybrid advances in those fields, also think it’s utter tech bro fantasy & fallacy to completely remove human factors from those societal constructs. Algorithmic biases are real, unjust & not all solvable, and likely hackable. #MapNotTerrain

Even w/ temporal versions of topo optimization/“AI”, allowing for constant readjustments, if applied to governance, good luck convincing people (esp. those w/ high mental inertia) of constant minute policy changes to readjust system & enforcing it peacefully/non-authoritarian

If anything, 2020 has shown just how messed up & futile this would be... Also, more generally what would algo optimize for? Optimization implies strong bias & loss of other aspects. E.g. in design most topo opt optimize for min weight/material. What does one optimize society for?

More importantly, who would/COULD be trusted w/ encoding the desired rules/contracts/policies, esp. if dynamic self-change is necessity for long-term success. Who watches watchmen? Also, gradient descent risks getting stuck in local minima. How to ensure to get out again? Etc...