A post from Gene Expression by Razib Khan who some of you may also know from the old gnxp site or perhaps from his BHTV debate with Eliezer. Some thoughts on the problem of trying to optimize your interactions to help you be less wrong. Your time is quite limited. Expect trade-offs.

A few days ago I was having drinks with some friends, and it came up that some of them had only recently become conscious of the fact that I leaned more toward the Republican party than the Democratic (I had remarked that my wife preferred that I keep my sideburns, as otherwise I would look too much like a Republican…though I sort of was one!). More shockingly for them was that I did not consider myself a liberal. I was somewhat bemused by the whole situation because it isn’t as if I’m particularly shy about expressing my various politically-incorrect opinions on any specific topic at work or play (these are people who I have met within the past ~2 years).

I assume that the problem here is that I violated a cognitive schema: liberal people are smarter than conservative people. Since I was conservative, they were, logically, smarter than me. The reality is probably not so convenient for the theory in this case, generating some dissonance. In the course of conversation I expressed frankly what I actually do hold to be a rough & ready approximation of my attitude toward discussion: I have almost no interest in persuading anyone of the truth of my particular views on any issue. This was relevant in that context because on occasion people try and draw me out as to the details of my disagreement with the consensus on an array of topics, when I often have no interest in expending the mental energy to do any such thing. It isn’t that I’m worried about getting into any argument with everyone else in the room. My friends are mostly natural scientists so I am very confident that I can alone hold my ground on any topic having to do with history and quantitative social science. Rather, the problem is my worry as to the point of it all. Who exactly is being edified by such exchanges? I never learn anything, as I am well acquainted with the standard arsenal of conventional Left-liberal talking points, while my interlocutors are often too amazed as my incomprehensible existence (i.e., not stupid, but not right-thinking) to really take in anything I’m saying.

Yet on a one-on-one basis I am much more likely to be open to a deep and thorough exchange. Why? The dynamic of signalling and group conformity is strongly dampened by removing third party observers from the interaction. With that tension removed I myself often feel less irritated if I have to invest a great deal of background information to make my own position clearer. Similarly, I often feel that my interlocutors are much less likely to trot out hackneyed and unpersuasive, but group approved, arguments.* There is quite often idiocy in crowds.

Ultimately we have to take a step back and reflect on what the point of it all is. For me the answer is rather easy: the point of it all is to understand the shape of reality as best as I can. It is impossible to do such a thing sitting back in an armchair and reflecting as an individual. Learning is a social process. You need feedback from others, and you need to mine and cull appropriate data and analyses from those who are more well versed in a given topic than you are. This is not easy, and time is finite. Avoiding stupid people is easy. The more difficult trick, at least for me, is avoiding smart people who offer stupid opinions on topics with which they are absolutely unfamiliar.** Creationist engineers are classic cases of the power of ignorance in the hands of the intelligent.

This brings me to learning more generally. Obviously I have no problem with people being autodidacts. Today the ability for one to be an autodidact has greatly expanded, but with power comes responsibility, and the necessity of prudence. I’m speaking obviously about the internet. But now we have the rise of online education. Recently MRUniversity opened, and Khan Academy is already rather famous. Tyler Cowen and Alex Taborrak’s endeavor has already received some praise:

MRU is ultimately aiming for a better actual education, not a better means of signaling. Cowen and Tabarrok are betting that there is an extraordinary amount of dead weight in current university classes (for example, on MRU the professor need not repeat himself as he inevitably must during live lectures, because if a student requires repetition, she can just watch the video again). “You can think of this,” Cowen says, laughing for the only time during our phone conversation and only lightly, “as a marginal attempt—a marginal revolution, so to speak—to get education to be more about learning.”

I am moderately skeptical, but I also think such experiments are necessary. Over the long term it seems likely that new forms of educational delivery and assessment with replace the middle and lower tiers of American higher education, and modify even the elite levels. But I don’t think we know yet what the exact nature of the information ecology is going to be.

Here is what I’d really like in the future: an app which analyzes someone’s stream of assertions and immediately assesses whether they are full of crap or not.*** There are many domains where I can do this analysis myself, and know to tune someone out because I know they’re signalling to ignorant people. But, there are many, many, more domains where I am ignorant and lost, and may fall prey to the bluffs and assertions of high caliber signalers, who have fashioned the simulacrum of intelligence. More concretely, people who are trying to impress without deep knowledge often fumble on many facts, something which could be run through an application such as WolframAlpha.

Of course things have changed a great deal. Over the past few years smartphones have cast a pall over the skills of the professional bullshitter. I think that there has been a qualitative change for the better.  Bullshitters known that they need to be cautious, so there is a preemptive effect.

* I am never in social circumstances where the political context is conservative.

** You also need to avoid socializing only with your own ideological set. This is easy for me since I don’t socialize with anyone who shares my politics or metaphysical opinions.

*** Looking things up manually is time consuming.

Debating group consensus with the group is less productive than debating it with individuals making up that group. Avoiding smart people who offer stupid opinions on topics with which they are absolutely unfamiliar is expensive. The internet has made this somewhat harder. We should like make an app to fix this or something.