My review of Inadequate Equilibria.
The most significant episodes in my career have been when I stood for heterodox beliefs. For that reason, Yudkowsky’s book raised issues that matter to me, even though I did not always find Inadequate Equilibria to be clear or convincing.
This essay proceeds as follows. I will articulate Yudkowsky’s two major themes in two ways, first using jargon from calculus and statistics, then using plain English. Next, I will tell some stories from my own life that relate to these themes.
Because this is an important topic for me, I recommend reading my entire essay.
Great note. Your cogent examples, starting with Emperor’s New Clothes, but including the “the heterodox management ideas of a low-level employee.” (sounds like me…), help illustrate the problem and issues.
I also like “local optimum”, but suspect that “inadequate equilibrium”, despite being longer and harder to say, might yet be better for marketing the idea. Most folks think they understand the current status quo, so calling it inadequate is a fine way to challenge them enough to be interesting but not quite threatening.
The plug for your book, Specialization and Trade, is well placed. This is a better name than the longer Patterns of Sustainable…, tho your intro does mention that the latter is included.
You refer to three kinds of terms:
1. Calculus: Local Optimum
2. Statistics: Excess of Type II errors regarding the null hypothesis.
3. Yudkowsky English: Inadequate Equilibrium.
I’d like to propose a borrowing from Economics: Social Failure (akin to ‘Market Failure’)
A market failure means there is an ‘inefficiency’, that is, there is a possible way to make things strictly better (for instance, by making at least one person better off while making no one else worse off), but for some reason the actions of agents seeking material self-interest in the marketplace lead to, and get stuck at, an inferior, inefficient outcome.
A social failure is an analogous situation when the actions of agents seeking social self-interest (i.e., status) get stuck in harmful and stubborn deviations from reality – often bordering on truly pathological levels of severity – and penalize any efforts at correction, reform, or improvement. Just as engineered structures or systems have failure modes, social institutions have social failure modes.
The big question is how to shake up a bad equilibrium so that one can open a path towards a superior outcome. I’d guess that often times there is really no good alternative but to delegitimate and significantly reduce the public trust and esteem given to a particular social institution. That is, tear down to rebuild better. Delegitimization comes with some significant risks and trade offs, however, so it isn’t always a good option.
A kind of example of a positive delegitimization of an institutional social failure is captured by the term “replication crisis”. Replication crisis stands for throwing decades of work of entire fields and disciplines into significant doubt, which is costly and probably throws some babies out with the bathwater. But it had to be done, for there to be any chance to start fresh and do better.
One can only hope that “local optimum” or “Excess Type II errors” or “inadequate equilbria” or “social failure” can similarly stand for a much more general proposition, which is that practically everything which humans discuss socially is susceptible to a number of predictable phenomena and processes whereby the results of those social interactions can remain unreliable and inaccurate for a long time even though all the top experts appear to share a consensus about it.
Hopefully too an approach which focuses on particular, common social failure modes that will help to identity and distinguish legitimate contrarian criticism from the much more numerous number of crackpots and cranks who dissent from mainstream views without any valid basis.
My personal favorite example of what looks to be a social failure: medical checklists
http://knowledge.wharton.upenn.edu/article/checking-everything-off-except-the-resistance-to-change/