The beginning of a new year is a customary time to take a look back and consider what has happened during the last 12 months. And while the time for doing so is admittedly rather arbitrary - after all, "years" do not really exist in the universe, just in our heads - it is useful and fun to review one's accomplishments every now and then. And a time when everyone else is doing it gives us a nice Schelling point for joining in, so we can pretend that it's not quite that arbitrary.

So what might be some noteworthy things that happened on Less Wrong in 2012 that could be worth mentioning?

Site upgrades

First, I would like to say "thank you" to all the people working on keeping this site running and helping it make increasingly more awesome! This obviously includes pretty much everyone who comments, posts and writes here, but particularly also the folks at Trikeapps, and everyone who contributes updates to the site's codebase. There were several site upgrades in 2012, four of which were major enough to get separate announcements:

Less Wrong's new front page was rolled out in March, thanks to work by matt. One can easily access a number of site features from the brain graphic, and there's a convenient introduction under it, together with links to featured articles and recent promoted articles. Hopefully, this has made it easier for newcomers to get familiar with the site.

The "Best" sorting system for comments was introduced in July. The work was done by John Simon, and integrated by Wes. Whereas the old default sorting system, "Top", favored old comments that had already floated to the top and were thus more likely to get even more upvotes, "Best" attempts to give newer comments a fairer chance.

In August we got the ability to show parent comments on /comments. The work was done by John Simon, and integrated by wmoore. This change makes it far easier to grasp the context of things seen on the recent comments page, given that we now see the old comment that the new comments are replying to.

And finally, starting from September, we have been able to write comments that contain polls! Work on the code was originally began by jimrandomh, finished later by John Simon, and deployed by wmoore and matt. Although people had long been taking advantage of comment vote counts as a crude way of creating their own polls, this change makes things far easier.

Meetup booklet

In June, we published the How to Run a Successful Less Wrong Meetup booklet, which I wrote together with lukeprog, and which got its graphical design from Stanislaw Boboryk. Numerous other people also helped, both by providing advice and by contributing pictures to it. In addition to general advice on running a meetup, it contains various games and exercises as well as case studies and examples from real meetup groups from around the world.

Index of original research

Starting from October, lukeprog has maintained a curated index of Less Wrong posts containing significant original research. It contains numerous posts, organized under categories such as general philosophy, decision theory / AI architectures / mathematical logic, ethics, and AI risk strategy. Last updated on December 17th, it now links to a total of 78 different posts.

Who are we?

In November and December, Yvain continued his hard work in holding the yearly survey. Among other interesting details, around 90% of us are male, 55% are from the USA, 41% are students and 31% are doing for-profit work. See the 2012 survey results for many more details.

Most popular posts of 2012

On LW, people tend to judge the popularity of a post by the number of upvotes that it has. But this only reflects the opinion of the registered users who care enough to vote. For purposes of this article, we were interested in finding out the posts that had made the biggest impact on the whole Internet. Although it's not a perfect measure either, we decided to measure popularity by the number of unique pageviews, as reported by Google Analytics.

Overall, in 2012 Less Wrong had over eight million unique pageviews and close to two million unique visitors (8,225,509 and 1,756,899, respectively). Of the posts that were written in 2012, the most popular ones were...

#10: Get Curious, in which lukeprog suggests that one of the most important rationality skills is being genuinely curious about things, instead of just jumping to the first answer that comes to your mind and leaving it at that. He suggests a three-step approach for actually becoming more curious: first, feel that you don't already know the answer, then start wanting to know the answer, and finally sprint headlong to reality. Together with a number of exercises intended to make you better at these steps, this article made a lot of folks curious about Less Wrong and caused people to sprint headlong to the post 10,850 times.

#9: Being curious about things means that you genuinely want to know the truth. That makes it useful to have a good grasp of The Useful Idea of Truth. This article by Eliezer Yudkowsky starts the Highly Advanced Epistemology 101 for Beginners sequence by explaining what exactly it means for something to be "true". In order to avoid spoiling the article's "meditations" for anyone who hasn't read it yet, I will not attempt to summarize the answer. I'll only suggest that one definition for "truth" could be the correctness of the claim that this post was viewed 11,161 times.

#8: Having defined truth, we can move on to ask, what are numbers? And in what sense is "2 + 2 = 4" meaningful or true? Eliezer Yudkowsky's Logical Pinpointing attempts to answer this question, partially through the cute device of conversing with an imaginary logician who understands logic perfectly but has no grasp of numbers. As they converse, they define the rules according to which arithmetic works. I'm going to skip the obvious pun due to it being too obvious, and only say that this article was viewed 12,606 times.

#7: Now that we're curious and understand both the meaning of truth and of numbers, it stands to reason that we should Be Happier than before. Or maybe not, since Klevador's article does not actually mention "understand obscure philosophy" as a way of getting happier. What it does mention is a big list of other things that have been shown to increase happiness. We first get a list of brief recommendations a few sentences long, and then somewhat longer excerpts of the relevant literature. There's also a full list of references. Let's hope that the 14,178 views that this post got made someone happier.

#6: Getting into more controversial territory, lukeprog advises us to Train Philosophers with Pearl and Kahneman, not Plato and Kant. Philosophy is getting increasingly diseased and irrelevant, he argues, and the cure for that involves incorporating more actual science and rationality into the standard philosopher curriculum. If the discussion on Hacker News is any indication, this post got a lot of people incensed, which might help explain why it got 14,334 views.

#5: Now that we got started on calling whole disciplines diseased, let's look at Diseased disciplines: the strange case of the inverted chart. Morendil's post begins with a hypothetical example of numerous academics all citing a particular source, which doesn't actually contain the intended reference... and then the intended source doesn't actually have the data to back up its claim, either. But that's just a hypothetical example, right? Well, not really, which helped this post get 17,385 views.

#4: Interestingly, our fourth-most-popular post isn't actually an original contribution as such. Grognor's transcript of Richard Feynman on Why Questions discusses the nature of explanations, and the fact that there are some things which simply cannot be adequately explained in terms of pre-existing knowledge. Instead, one has to learn entirely new concepts in order to comprehend them. Hopefully, at least this much was understood on the 18,402 times that the post was viewed.

#3: From physics to neuroscience: kalla724's Attention control is critical for changing/increasing/altering motivation explores the effect of attention on neural plasticity, including the plasticity of motivation. It explains that paying attention to something can increase the amount of brain circuitry dedicated to processing that something, generally by repurposing nearby less-used circuitry. This also has practical applications, such as in helping to explain why Cognitive Behavioral Therapy works. That earned the post 21,136 views.

#2: I should be writing this post instead of browsing Facebook. Fortunately, lukeprog has a post titled My Algorithm for Beating Procrastination. Based on the equation of Motivation = (Expectancy * Value) / (Impulsiveness * Delay), the algorithm involves first noticing that you are procrastinating, then guessing which part of the motivation equation is causing you the most trouble, and then trying several methods for attacking that specific problem. I guess that a lot of people shared this on Facebook where other procrastinators saw it, because the article got 38,637 views.

#1: And finally... the most read 2012 article on the site was Yvain's The noncentral fallacy - the worst argument in the world?, where he defined the noncentral fallacy as "X is in a category whose archetypal member gives us a certain emotional reaction. Therefore, we should apply that emotional reaction to X, even though it is not a central category member." Which sounds pretty abstract, but the political examples in the post should make it clearer. The politics probably helped contribute to this post's achievement of 41,932 views.

Most popular all-time posts

In addition to looking at only the posts that were made in 2012, people might be interested in knowing which posts were the most viewed in 2012 overall. The top three ones were all written by lukeprog, and we can see that two of them were closely related to the top-scorers which were written last year.

How to be Happy is LW's run-away favorite article and was viewed more than every page on LW except the home page and the discussion homepage. That is, 228,747 times! The Best Textbooks on Every Subject comes as a distant second at 98,011 views. And the third one is How to Beat Procrastination, at 66,587 views.

So I guess the take-home message is: people want to be happier, smarter, and more productive. Let's keep becoming those things in 2013!

New to LessWrong?

New Comment
35 comments, sorted by Click to highlight new comments since: Today at 1:53 PM

So, on Reddit I notice that asking after the "most popular" posts usually brings up the least intellectually stimulating content, presumably because easy content is processed and upvoted faster. Even in a high quality subreddit, clicking the "top all time" icon brings up memes and vapid statements.

Now, while I have to say that all these Lesswrong posts were definitely high quality and worth my time to read, I do think that among these, the most popular posts and articles were not among the most thought provoking ones that we have here.

Instead, the most popular posts and articles here basically restate statements which I already understand and rather strongly agree with. Yes, I think they are insightful, but that is primarily because they mirror my own pre-formed views on the matters being discussed and like most people I tend to find my own opinions insightful. Whereas, those posts which actually gave me a novel idea to play with were not necessarily the most popular.

Note that this is not a criticism against the top articles themselves (I enjoyed reading them immensely). I'm just pointing out a trait inherent in user-generated moderation (up/down votes). The top content tend to be shallower and more agreeable, whereas more substantial or controversial content tends not to float to the top as quickly.

I think the reason that this happens (in general) is that shallow content can be processed quickly by the reader, and it is easy to judge whether or not you like it. With more substantial content, it takes a while to process and it is more difficult to decide whether or not you like it because it is more sophisticated. So even though nobody really likes reddit internet memes, they float to the top because they are good at getting votes.

What's interesting is that lesswrong users love substantial content and set quite a high standard. The users here will downvote anything which is too shallow. So our Lesswrong variant on the "internet meme" (uncontroversial, easy to follow generalizations that we all agree with) are still very well written and a pleasure to read.

It's just an interesting dynamic and I thought I'd point it out.

I think of these articles, the only one which really taught me something was "logical pinpointing" (explaining first and second order logic, and introducing the viewpoint that axioms as "pin down" concepts rather than construct them) but that was all the way at #8.

That's not to say it was the best article here of course - just that I personally was either already familiar with or had independently arrived at the ideas described within the others. That particular article's novelty is probably peculiar to me (quite a few of the other articles would have been novel to me four years ago, for example) but I suspect that if everyone here were to list which articles on lesswrong taught them something they didn't already know, only a small portion of that list would end up on top (once adjusting for the fact that top articles are viewed more often)

What if our voting system looked something like this, making room for both the enticing subjective vote as well as the democratic objective one? That tweak would fix all my bad voting habits.

I get the first three (fun to read upvote, novel or educational upvote, and downvote)

but what's the last button mean?

"I agree to follow this course of action or policy and / or I think this person's opinion is correct?"

And how would we explain to new users that it isn't what it looks like...also, I'm wondering what context this image was originally created in...

On a serious note, we don't want to get too complicated...but I wouldn't object to a three button system. "Like" "Dislike" and "I personally learned something new, considered a novel idea, or updated an opinion as a result of this article or comment".

Because while I tend to like reading well written articles which phrase my opinions more articulately than I myself could, I'd also like to be able to bring up the more difficult and rewarding pieces on command.

I'm guessing since we've apparently established that the average user here is gifted, we can trust people to use the "novel/learn/update" button correctly and with appropriate frugality. It would also be good for feedback purposes.

Seeing how you would interpret each icon based on context was an experiment in design.

The last icon is for the "circlejerk": your objective sad face.

This is the part of my reply where i see i should offer more time to collecting hyperlinks for examples, but i hope it follows easily enough:

Relevant xkcd to voting systems: happy button only from me.

Someone saying they got their bike stolen while they were voting: sad button only

Peer-reviewed paper about voting in forums: brain button only

Peer-reviewed paper about the shape of Antarctica: sperm button

Peer-reviewed paper about Maru and the evolution of voting on Youtube: happy button and brain button from me.

Pun threads concerning voting: happy button and the sperm button from me.

Any opinion shared with Caps Lock: sad button and sperm button

Peer-reviewed paper about gun violence and voting systems: sad button and brain button from me.

Any other combinations would defeat themselves, so im don't see any reason to click three buttons or all four. Perhaps the objective votes of brain and sperm are unlocked the way voting is unlocked in AskUbuntu's forums: only after certain contributions and/or circumstances are met.

That's what i'll be fighting for at least. I vote subjectively and i vote objectively. If you make me choose, i'm going for the subjective vote every time. Best to break the two apart.

sperm: Brains have two hemispheres; i wonder if there's correlation....

Polls in comments was also a pretty big feature added this year.

Damn, how did I forget that? Thanks, I added a mention of it.

To counteract the self-congratulatory tone of the article a little bit, here are some low points, in my observation:

  • The site is often unbearably slow and occasionally times out. I suppose this is a side effect of its popularity combined with a shoe-string budget.

  • The site upgrades are basically useless for me. Specifically, the front page needlessly duplicates the only useful link on it (discussion/new) and does not have the other useful ones (comments and recentposts) except on the sidebar. I played with the "Best" sorting and ended up reverting to "New". I played with the "show parents" setting, but it ends up cluttered with duplicate comments, so I had to turn that off, too. Maybe others find these setting more useful.

  • The anti-troll measure often hides controversial but clearly non-trolling comments from respected regulars (the proposal to exempt those with a respectable 30-day karma from being labeled as trolls was declined by EY).

I'm also wondering if maybe it would be a good idea to have top-10 all time/last year/last month posts and comments featured on the front page.

The site is often unbearably slow and occasionally times out. I suppose this is a side effect of its popularity combined with a shoe-string budget.

I don't think I've ever noticed that. Maybe it's slower when lots of people use it at once, and I live in a time zone different from that of most LWers.

The most upvoted post of all time on LW is Holden's criticism of SI. How many pageviews has that gotten?

Google Analytics says that it got 9,334 unique pageviews in 2012. (Though for some reason GA also seems to me giving slightly different numbers than it gave me when I composed the post yesterday, so I'm not sure if they're completely reliable? But the difference is pretty slight, doesn't change the relative rankings of the different posts.)


"Posts most popular in 2012 and written in 2012" gives an advantage to posts written earlier in the year. Solutions to this could include:

  • rank posts by (2012 views / age), though that is noisy at the recent end
  • rank posts by the number of views they had when one year old
  • rank posts by the "Best" comment sorting algorithm; wait, that doesn't work, there is no analogue of a downvote, so make this "use a more refined estimation algorithm"

The link for Feynman's Why Questions is broken.

Fixed, thanks.

Nope, it's still broken.

Damnit. Fixed again, hopefully for real this time.

I have just read the Hacker News discussion of Luke's Train Philosophers with Pearl and Kahneman, not Plato and Kant. Ouch.

One thing that strikes me there (and that I have observed elsewhere), is the inability of many people to separate philosophy from its history. They assert that to hope to have more than a surface understanding, you must read all those long dead people. This is certainly true when one does original historical research, but completely useless in most other situations, where it is more productive to seek curated input.

Most other fields are much more reasonable. For instance, nobody in their right mind would try to teach Newtonian mechanics with Philosophiæ Naturalis Principia Mathematica. This book is just no longer relevant to the teaching and advancement of modern science. Newtonian mechanics are not exactly true to begin with, and we now have better material than the original book to study it. We could see that as a criticism of Newton, but really, that's a praise. He advanced Science so well that others were able to build on his foundations, and advance even further!

This should be as obvious as causal graphs: When A causes B and B causes C, the knowledge of A is completely irrelevant to C as soon as we know B. Conversely, when old work has been captured, assimilated, refined, and transformed into brutally efficient mechanisms of learning, there is no need to refer to them any more. Why this line of thinking doesn't seem to apply to philosophy is beyond me.

Well, to be somewhat fair, several of the commenters there are not saying "read original sources for the sake of reading original sources" but rather "read original sources because they have value which this particular branch which developed from it has lost sight of." Which is not as obviously wrong as a matter of logic as you make it sound. The claim is precisely that in addition to B and C, there is a potential B' which could be derived from A, and that B does not include the elements of A from which B' could be derived.

That said, it seems the proper response to that claim is "what value, and why should I believe that it's worth the effort involved in obtaining it?"

Hmm, maybe my pattern matching was too quick. My prior for "Philosophy and history of philosophy are conflated" is quite high, since most philosophy courses in French high schools basically teach us what old dead guys wrote on things like Love, Death etc. I did have an actual philosophy professor for 2 months though (he tried to teach semantic and logic).

By the way, I also strongly suspect that Philosophy is often also treated as a literary discipline (at least here in France). The two biggest clues are the obligation to read the original works, and the refusal to admit that such works could be correct or incorrect as a simple matter of fact¹.

(1) In our high school final exam, any definite answer to the philosophy question is shot down. We are mainly supposed to spit back old dead wisdom, and conclude somehow that we don't have the answer. Even if the question is as silly as "Can opinions at odds with the facts be true?"

I guess this is one of my problems with philosophy: unlike in sciences, there is no B, only a collection of disjoint and often incompatible B's with various asterisks.


(Disclaimer: I am not very familiar with how Philosophy or Physics are usually taught.)

Physics students begin by studying Newtonian mechanics. Teaching has a form different from the original, but the mathematical formalisms remain. Philosophy students begin by studying Aristotelian logic (or whatever). If philosophy were taught without original work, would Luke's criticism not hold? Isn't that already largely the case?

Not that I disagree with the conclusion, but the argument can be turned around remarking that obsolete formalisms are useful toys in both disciplines.

Now I wonder if teaching NM is detrimental to Physics research...

Newtonian mechanics is not obsolete. It is just founded on assumptions that sometimes break down, but sometimes they don't break down. There's still plenty of stuff you can do with Newtonian mechanics, e.g. build bridges. And any future theory of physics is constrained by the fact that in certain limits it has to reduce to Newtonian mechanics, so it still serves to substantially control what future physics can look like.

This is not true of obsolete formalisms in philosophy, which might be founded on fundamental confusions about how words work or whatever.


Tangent to the discussion, this is interesting to me:

any future theory of physics is constrained by the fact that in certain limits it has to reduce to Newtonian mechanics

What is the reason for this? what limits?

Newtonian mechanics still correctly predicts the behavior of things that are big and slow, e.g. bowling balls. Therefore, in the limit as things become big and slow, any future theory of physics has to reduce to Newtonian mechanics.

Alternatively and more specifically, special relativity is constrained by the fact that it has to reduce to Newtonian mechanics in the limit as the speed of light tends to infinity, and quantum mechanics is constrained by the fact that it has to reduce to Newtonian mechanics in the limit as Planck's constant tends to zero. Physicists call this idea taking the classical limit.


Conceded, NM have applications beyond its use as a learning tool. But that should be considered relative to the discipline as whole.

The usefulness of ancient philosophy (very low) relative to modern philosophy (low) seems comparable to that of ancient physics (medium) to modern physics (high). The assessment I am least certain of here is that of modern philosophy.

I enjoyed the post, but I'm confused by the first paragraph. Does it add to the post to have an explicit justification?

I observe that some people benefit from short introductions to get their consciousness (or stream-of-consciousness or whatever) into the right place within their own mental space and be able to run and process what they're reading at a better pace.

For some subjects this is certainly the case for me - for instance, formal mathematics seems to be in a relatively separate mental / memory cluster in my mind that is rather far from other clusters, and takes a bit more priming for me to get into, while other people seem to comparatively be instantly "there" and understand what is going on even when you spring mathematical equations at the edge of their understanding on them while they were wondering what socks to wear.

So I don't feel that it adds to the post's content, but I feel that it is still potentially useful for some readers in order to better absorb and more easily process said content.

Probably not, but I felt that it needed some introduction. And I'm bad at writing introductions.

And I'm bad at writing introductions.

Hardly; you already wrote a fine introduction:

First, I would like to say "thank you" to all the people working on keeping this site running and helping it make increasingly more awesome! This obviously includes pretty much everyone who comments, posts and writes here, but particularly also the folks at Trikeapps, and everyone who contributes updates to the site's codebase.

Then add on something like "Here are some particularly noteworthy things that happened in 2012:"

Only on less wrong is a new years retrospective justified using game theory ;)

I chuckled slightly at the "schelling point" line and thought the introduction was fine.


And I'm bad at writing introductions.

Hardly; you already wrote a fine introduction:

First, I would like to say "thank you" to all the people working on keeping this site running and helping it make increasingly more awesome! This obviously includes pretty much everyone who comments, posts and writes here, but particularly also the folks at Trikeapps, and everyone who contributes updates to the site's codebase.

Then add on something like "Here are some particularly noteworthy things that happened in 2012:"

[This comment is no longer endorsed by its author]Reply

So I guess the take-home message is: people want to be happier, smarter, and more productive.

I, uh, can't read that without hearing this.

Were there some Freakonomish economists around, it would be good to find out which peculiar and seemingly unrelated factors correlate with most viewed posts. I'll bet there is some mild correlation between skull size and number of viewers, even if you disregard Lukeprog's posts. Disentangling the variables would be probably unfeasible though.

I wonder about the list of "Would be most visited LessWrong articles if ever written"

Now planning on writing at least two pages per day in 2013... some will end up in LessWrong I suppose. Any LWers with a new years resolution related to LessWrong, or being more rationally productive in general?

I aim to spend at least 2 hours a week learning to program.