That’s a favorite of mine. Written right after the attacks. But I wonder why he delisted it while still hosting the essay to make it hard to find but still accessible
https://www.paulgraham.com/hijack.html
Paul Graham compares the hijacking of an airplane to a buffer overflow attack.
He doesn't claim that tools for writing more secure computer programs could prevent terrorists from hijacking a plane.
He proposes the idea that some similar reasoning can be applied to preventing hijackings. It's not enough to monitor who gets on the airplane carrying what, but to separate the cockpit space from the passenger space.
I don't think his point is very good because array overruns in a heap can be exploitable. Lisps (and other kinds of run-times) don't necessarily separate code from data in their heaps.
The best kind of mean for todays environment seems to be Machiavellian coated in a superficial gloss of magnanimity via stuff like effective altruism, longtermism, or wokewashing.
My thesis is that malicious behavior by executives is in some large percentage of cases not intentional but uncontrolled profit-seeking behavior by the ant-heap that is their organization.
Otherwise you have the "Just following orders" excuse.
This isn't an academic issue. Corporate harms are pervasive, very real, and hugely destructive to the economy and to the well-being of (at least) hundreds of millions.