Most new conjectures are wrong. Yes good. Most old conjectures are wrong. Also fine. This is because new and old are pretty much unbiased subsets of conjectures in general. However the set of ideas in active use are a biased, selected set. Imperfectly selected yes, and quite slow, yes. And can cause a lot of misery, sure. But implemented laws are a selected set and so have a higher percentage of true ideas than the set of new and the set of old.
There’s a lovely power law at work here, a kind of self-similar Sturgeon’s law. Most governments are static except a few, and within those few, most states are static except a few, and within those states most people are static except a few, and within each person most ideas are static except a few. We are utterly blessed \footnote(actually by the hard work of some of our ancestors) with a system of government that, however slowly, however imperfectly, manages to asymmetrically select good ideas, eeking out progress in this sliver of enlightenment. I would say that western democracies are more effective and efficient as any other system ever for social change (that also survived the changes), I would think you would too. It’s the worst method of selecting good ideas, except all the others so far.
You mentioned a few discernible mechanisms that yet undermine selection in our system. But I really can’t be sure that they do. “Gut reactions”? How about human intuition. “Campaign goals”? Um… responding to the electorate. I know its tenuous, but the only real way to tell how damaging or helpful these mechanisms are is by looking at the results, and comparing to the alternatives. On the local scale the political representatives could be as dumb as ants, I don’t care, so long as progress emerges.
Actually at the last second I think I might be giving myself to much credit. If it turned out autocracy truly had the best outcomes, I wouldn’t be comfortable with it. Nor would submitting to representation by mutant ant-people. And here I feel like I see the outline of the problem. Such a complex system means that it’s easy to pair every success of a system with a mechanism one likes, and problems in a system with the lack of something one wants. In reality, given some person’s beliefs, we should expect them to like some parts of government that actually don’t work, and to hate parts that do.
So the only thing I can be sure of is that the best government is probably not the one I think would be best. And certainly not the one I hope would be best.
Deutsch and Taleb avoid the above problems by ignoring detailed mechanics and focusing on results. But they do it from different angles. Taleb looks back, the ancestors he worships had strong results on the survival front, with city states that lasted millennia. To stand up for them is to say: because they survived longer, they have mechanisms we need to implement. If we need an autocracy, we should use an autocracy. But Deutsch with reference to the present results looks forward. There’s a kind of ancestor ambivalence generated from his cutthroat definition of sustainability: Is the society around now? No? Pfft, too bad. It failed because it didn’t progress enough. Will we eventually find that Automaton ant-people give us the best results? Fine. We’re still here, and the only way to guarantee survival is to ride this sliver to infinity.