Chesterton-Miller Behavioral Fence: when you see people behaving in a seemingly suboptimal way, ask what you’re missing about what they’re actually optimizing for.
Behavior is rarely suboptimal, but assumed cost functions are usually wrong.
Conversation
Replying to
Apparently it needs to be explicitly stated that most people’s cost functions are mostly illegible, most of the time. We’re talking unconscious optimization and equilibria here.
3
2
57
Corollary: when someone leaves out obvious qualifiers on a complex proposition, it’s generally safe to assume they’re solving for 240 characters and trusting readers to fill in the obvious supporting clauses
2
10
115
Good question. The theory does apply to itself consistently.
Usually when people assume the wrong cost function, they’re optimizing for power, influence, and control
Quote Tweet
Replying to @vgr
When I assume the wrong cost function, what am I really optimizing for?
1
2
43
Or conversely, reflect on one’s own unconscious authoritarian tendencies 🤣
Quote Tweet
Replying to @provisionalidea and @vgr
In addition to helping make sense of other cultures and worldviews, it turns out to be a brilliant way to identify perverse incentives and systemic inefficiencies.
1
1
21
Show replies
Replying to
Or the cost function is externalized (ie they are not on the hook for what the full costs should be)
4
Replying to
Fences are never built for no reason. Irrationality can exist for no (good) reason.
1
4






