Reply to: Extreme Rationality: It's Not That Great
Belaboring of: Rational Me Or We?
Related to: A Sense That More Is Possible
The success of Yvain's post threw me off completely. My experience has been opposite to what he describes: x-rationality, which I've been working on since the mid-to-late nineties, has been centrally important to successses I've had in business and family life. Yet the LessWrong community, which I greatly respect, broadly endorsed Yvain's argument that:
There seems to me to be approximately zero empirical evidence that x-rationality has a large effect on your practical success, and some anecdotal empirical evidence against it.
So that left me pondering what's different in my experience. I've been working on these things longer than most, and am more skilled than many, but that seemed unlikely to be the key.
The difference, I now think, is that I've been lucky enough to spend huge amounts of time in deeply rationalist organizations and groups--the companies I've worked at, my marriage, my circle of friends.
And rational groups kick ass.
An individual can unpack free will or figure out that the Copenhagen interpretation is nonsense. But I agree with Yvain that in a lonely rationalist's individual life, the extra oomph of x-rationality may well be drowned in the noise of all the other factors of success and failure.
But groups! Groups magnify the importance of rational thinking tremendously:
- Whereas a rational individual is still limited by her individual intelligence, creativity, and charisma, a rational group can promote the single best idea, leader, or method out of hundreds or thousands or millions.
- Groups have powerful feedback loops; small dysfunctions can grow into disaster by repeated reflection, and small positives can cascade into massive success.
- In a particularly powerful feedback process, groups can select for and promote exceptional members.
- Groups can establish rules/norms/patterns that 1) directly improve members and 2) counteract members' weaknesses.
- Groups often operate in spaces where small differences are crucial. Companies with slightly better risk management are currently preparing to dominate the financial space. Countries with slightly more rational systems have generated the 0.5% of extra annual growth that leads, over centuries, to dramatically improved ways of life. Even in family life, a bit more rationality can easily be the difference between gradual divergence and gradual convergence.
And we're not even talking about the extra power of x-rationality. Imagine a couple that truly understood Aumann, a company that grokked the Planning Fallacy, a polity that consistently tried Pulling the Rope Sideways.
When it comes to groups--sized from two to a billion--Yvain couldn't be more wrong.
Update: Orthonormal points out that I don't provide many concrete examples; I only link to three above. I'll try to put more here as I think of them:
- In Better, Atul Gawande talks about ways in which some groups of doctors have dramatically improved by becoming more group-rational, including OB standbys like keeping score and having sensitive discussions privately (and thus more openly).
- Google seems like an extremely rational place for a public company. Two strong signals are that they are extremely data-driven and have used prediction markets. To be painfully clear, I'm not claiming that Google's success is due to the use of prediction markets, merely that these datapoints help demonstrate Google's overall rationality.
- As AlanCrowe points out in the comments, Warren Buffett and Charlie Munger have a rationalist approach.
It is true that groups magnify the importance of rational thinking, but I don't think the few examples cited prove that non-x rational thinking was the cause of the success and even if that were proven, it wouldn't prove that x-rational thinking would (given opportunity cost) make a group better. The first example is Toyota, but the link isn't a serious argument that the main cause for their success versus the US auto industry is the continuous improvement, nor that their adoption of continuous improvement was them being rational and the US auto industry not adopting it was them being less rational. I doubt all those claims.
I also doubt that "Companies with slightly better risk management are currently preparing to dominate the financial space." As best I can tell, this presupposes that the financial crises hit all companies the same and the difference in effects was due to risk management policies. I tend to think that the different crises hit different companies differently because the crisis were varied and somewhat random in their effects, so that some companies got lucky (though some probably did have better risk policies).
And I really don't understand the claim about certain countries having more rational systems which over centuries led to their current improved "ways of life." I assume this refers to the highly industrialized countries. Was their growth only .5% more per year than other countries? Was the Tsarist rule of Russia more rational than that of imperial China? Or Great Britain more rational than Denmark? I highly doubt that amounts of rationality is what explains the current economic distinctions between countries.
Finally, there is the cliche about finance professors--if they really understood it, why aren't they rich? Similarly, why not start a company using x-rationality to guide it? Find, for example, a new small business that is growing, and just do the same thing--only using x-rationality.
From the OP: "Whereas a rational individual is still limited by her individual intelligence, creativity, and charisma, a rational group can promote the single best idea, leader, or method out of hundreds or thousands or millions."
GIGO. At any given time, the rational group will be limited by their consensus beliefs and mental models. (Indicator of group quality -- do their mental models improve over time?) Rationality is just a tool for uncovering truth. I've often found that I'll go to work on a hard problem using rigorous intellectual tools... (read more)