It is a very interesting set of case studies and thank you for digging in the historical documents!
Several thoughts:
tradeoffs—speed vs safety (but just 5 deaths during Empire State Building construction, zero in Chrysler building; 60 people died in WTC construction, while Brooklyn Bridge construction cost the lives of 5% of workforce)
robustness—peer review in science takes a lot of time, but ensures that things actually work; it is also correlated with the difference between public and private control; nicely portrayed in “Don’t look up”
The problem: your predator-prey system leads by default to extinction.
Lotka-Volterra equations did not. Maybe something else should be changed in the model (like the carrying capacity? reproduction rate coefficient with prey abundance? the random mating simulation?), so that this problem is resolved in a similar way?
I know anecdotally, that models of evolving food webs typically are unstable. It would be great to hear more insights on this from the community :)