The Problem With Rational Wiki

Perhaps RW is more relevant than I thought. Rachael Briggs announced via an edit that she isn't doing the TDT paper for SIAI any more. (XiXiDu emailed and verified it was her.)

It strikes me that SIAI/LW may be at the stage of needing a proper PR strategy and someone whose job that is.

Briggs decided not to spend further time writing the TDT paper. However, SI is now paying her hourly to give us feedback on the TDT paper that Alex Altair is developing. She's very good at that, and appears to be enjoying the process.

2wedrifid7yIt is relevant enough that if people google their own name mentions therein can appear in the search results.
4Jayson_Virissimo7yAccording to XiXiDu, "she believed it would be unlikely for her to produce an article that would be satisfactory to both her and SIAI". Does that mean she didn't think it could be formalized, that it could be but that it would end up being obviously inferior to CDT/EDT or otherwise a bad decision theory, that she isn't able to do it but that someone else could, or something else entirely?

The Problem With Rational Wiki

by [anonymous] 1 min read26th Oct 201264 comments


Related to: RationalWiki's take on LW, David Gerard's Comments, Vladimir_M's comments, Public Drafts

I wanted to bring more attention to this argument because I've ran into related discussion several times in the comment section and because it demonstrates a failure mode that LessWrong may find itself vulnerable to.

Since it has been cited as a source especially on the reputation LessWrong may or may not have elsewhere I think readers should be aware Rational Wiki has a certain reputation here as well. I'm not talking about the object level disagreements such as cryonics, existential risk, many-worlds interpretation and artificial intelligence because we have some reasonable disagreement on those here as well. Even its cheeky tone while not helping its stated goals can be amusing. I'm somewhat less forgiving about their casual approach to epistemology and their vulnerability to cargo cult science, as long as it is peer reviewed cargo cult science.

While factually it is as about as accurate as Wikipedia, it is very selective about the facts that it is interested in. For example what would you expect from a site calling itself "Rational Wiki" to have on its page about charity. Do you expect information on how much good charity actually does? What kinds of charities do not do what they say on the label? How to avoid getting misled? The ethics of charity? The psychology, sociology or economics of charity?

I'm sorry to disappoint you but the article consists of some haphazardly arranged facts and stats on how much members of some religions give or are supposed to give to charity, a dig against Christianity and a non-sequitur unfavourable comparison of the US to Sweden. Contrast this with what you can find on the topic on sites like LessWrong or 80, 000 Hours. Basically the material presented is what a slightly left of centre atheist needs to win an internet debate. As is much of the rest of the site.

Indeed some entries have a clear ideological bias that is quite startling to behold on a "rational wiki" and it has been noted by some.

Now to avoid any misunderstandings there are good articles, a few LWers are contributors to the rational wiki and there is certainly nothing wrong with being a left of centre atheist! Nearly everyone on this site is an atheist, and people who identify as left wing politically form a large majority here. The tribal markers and its political agenda aren't the biggest problem. Sites with all sorts of agendas, even political ones, promoting rationality are a good thing.

Its problem is that it is an ammunition depot to aid in winning debates. Very specific kinds of debates too. This may sound harsh, but consider: How many people reading the site that aren't already atheists will change their mind on religion? How many people who follow a "crankish" belief won't do so afterwards? While I'm sure it happens the site obviously isn't optimized for this. How many people will read the wiki and try to find errors and biases in their own thinking to debug it instead of breaking if further with confirmation bias or using it as a club? How many will apply this knowledge to help them with any real world problems? Truth seeeking? As a source or community that could aid in that quest it is less useful and reliable than Wikipedia, which while a rather good and extensive encyclopaedia (despite snickering to the contrary) has a subtly but importantly different stated goal.

What else remains? What other plausible function does it serve?