MIRI needs volunteers to proofread our soon-to-be-released publications, such as Eliezer's "Intelligence Explosion Microeconomics." Some reasons to get involved:

• Get a sneak peek at our publications before they become publicly available.
• Earn points at MIRIvolunteers.org, our online volunteer system that runs on Youtopia. (Even if you're not interested in the points, tracking your time through Youtopia helps us manage and quantify the volunteer proofreading effort.)
• Having polished and well-written publications is of high-value to MIRI.
• Help speed up our publication process. Proofreading is currently our biggest bottle-neck.

Some of the papers that are sitting in the pipeline and ready for proofreading right now (or will be very soon):

• "Avoiding Unintended AI Behaviors" by Bill Hibbard
• "Decision Support for Safe AI Design" by Bill Hibbard
• "A Comparison of Decision Algorithms on Newcomblike Problems" by Alex Altair
• "Intelligence Explosion Microeconomics" by Eliezer Yudkowsky

How proofreading works:

• Youtopia, with the help of some shared Google Docs, is used to manage and track the available documents and who's proofread what.
• Proofreading entails checking for basic grammar, spelling, punctuation, and formatting errors; pointing out areas of confusion or concern; and making general style and flow suggestions.
• Don't worry, you don't have to proofread entire documents, just as many individual pages as you like.
• (This is explained in more detail once you've joined the MIRI Proofreaders group.)

How to join Youtopia and specifically the MIRI Proofreaders group:

1. Go go MIRIvolunteers.org.
2. In the right sidebar click on Register as a Volunteer and fill out your info.
3. Once your Youtopia account is created (this could take a day or two), head here and click on "Join Organization.”
4. Once your membership is approved you will have have access to detailed proofreading instructions and draft versions of our publications.

Questions can be directed to alexv@intelligence.org.

1. Don't.

This isn't really a full post, but merely a note of potential interest. Paul Graham (who runs Hacker News) has several very interesting and thought-provoking essays located on his personal website. To me they fit very well with the style of thinking employed and advocated by many people on LW and I'd advise that nearly anyone interested in LW check out his work.

I especially recommend Keep Your Identity Small, What You Can't Say, and What You'll Wish You'd Known, but nearly every essay up there is interesting to me in some way. Many of them are directly relevant to issues of rationality, while others are only indirectly related, but either way I found them worth my time.

https://plus.google.com/109794669788083578017/posts/gLgSnkCtgrR

Brain function relies on communication between large populations of neurons across multiple brain areas, a full understanding of which would require knowledge of the time-varying activity of all neurons in the central nervous system. Here we use light-sheet microscopy to record activity, reported through the genetically encoded calcium indicator GCaMP5G, from the entire volume of the brain of the larval zebrafish in vivo at 0.8 Hz, capturing more than 80% of all neurons at single-cell resolution. Demonstrating how this technique can be used to reveal functionally defined circuits across the brain, we identify two populations of neurons with correlated activity patterns. One circuit consists of hindbrain neurons functionally coupled to spinal cord neuropil. The other consists of an anatomically symmetric population in the anterior hindbrain, with activity in the left and right halves oscillating in antiphase, on a timescale of 20 s, and coupled to equally slow oscillations in the inferior olive.

Page down at the link to see the animation.

Today's post, Raising the Sanity Waterline was originally published on 12 March 2009. A summary (taken from the LW wiki):

Behind every particular failure of social rationality is a larger and more general failure of social rationality; even if all religious content were deleted tomorrow from all human minds, the larger failures that permit religion would still be present. Religion may serve the function of an asphyxiated canary in a coal mine - getting rid of the canary doesn't get rid of the gas. Even a complete social victory for atheism would only be the beginning of the real work of rationalists. What could you teach people without ever explicitly mentioning religion, that would raise their general epistemic waterline to the point that religion went underwater?

Discuss the post here (rather than in the comments to the original post).

This post is part of the Rerunning the Sequences series, where we'll be going through Eliezer Yudkowsky's old posts in order so that people who are interested can (re-)read and discuss them. The previous post was Striving to Accept, and you can use the sequence_reruns tag or rss feed to follow the rest of the series.

Sequence reruns are a community-driven effort. You can participate by re-reading the sequence post, discussing it here, posting the next day's sequence reruns post, or summarizing forthcoming articles on the wiki. Go here for more details, or to have meta discussions about the Rerunning the Sequences series.

I have often benefited from recommendations for Things I Didn't Know I Wanted.

Given that Less Wrong is a community of unusually intelligent, critical, and self-improvement-focused people, I suspect we can generate a pretty helpful thread of product recommendations — perhaps even a monthly thread of product recommendations.

Rules:

• Post one product your recommend per comment, so they can be discussed and voted on independently.
• Provide a link for purchasing the product.
• No books, movies, TV, games, or music. (These should go in other threads, like this one or this one.)

A fun little exercise in dissolving a problem. Relatively quick, but it can wake you up on a slow day.

http://www.memedroid.com/share-meme/337837/8000

Let's consider the agent given in A model of UDT with a halting oracle. One will notice that that agent is not quite well defined because it doesn't tell us in what order we are supposed to consider actions in step 1. But surely that doesn't matter, right? Wrong.

Let's consider the prisoner dilemma with payment matrix given by

and consider agent A which consider whether there is a proof that A()≠D before considering whether there is a proof that A()≠C and agent A' which do things in the opposite order. If A or A' is pitted against itself everything is well and mutual cooperation is the result of the game but what if A is pitted against A'? Then A break down and cry.

Let's call the utility functions of A U and the utility function of A' U' and consider a model of PA in which PA is inconsistent (such a model must exist if PA is consistent). In such a model we will have A()=D and A'()=C and so U()=5 and U'()=0. That means that A will not be able to prove that A()=D => U()=u for any u different from 5 and so either A will defect and A' will cooperate or A will break down and cry, but A' will not cooperate because it cannot prove A'()=C => U()=u' for any u' except possibly 0, so A will break down and cry. QED

More generally if M is a model of PA in which PA is inconsistent, an agent defined in this way will never be able to prove that A()=a => U()=u (where a is the first action considered in step 1) except possibly for u=u₀ where u₀ is the value of U() in M. That seems to create a huge problem for that approach to UDT.

I looked for Information Diet in Lesswrong search, and found something amazing:

On Lukeprog's Q and A as the new executive director, he was asked:

What is your information diet like? (I mean other than when you engage in focused learning.) Do you regulate it, or do you just let it happen naturally?

By that I mean things like:

• Do you have a reading schedule (e.g. X hours daily)?
• Do you follow the news, or try to avoid information with a short shelf-life?
• Do you significantly limit yourself with certain materials (e.g. fun stuff) to focus on higher priorities?
• In the end, what is the makeup of the diet?
• Etc.

To which he responded:

• I do not regulate my information diet.
• I do not have a reading schedule.
• I do not follow the news.
• I haven't read fiction in years. This is not because I'm avoiding "fun stuff," but because my brain complains when I'm reading fiction. I can't even read HPMOR. I don't need to consciously "limit" my consumption of "fun stuff" because reading scientific review articles on subjects I'm researching and writing about is the fun stuff.
• What I'm trying to learn at this moment almost entirely dictates my reading habits.
• The only thing beyond this scope is my RSS feed, which I skim through in about 15 minutes per day.

Whatever was the case back then, I'll bet is not anymore. No one with assistants and such a workload should be let adrift like that.

Citizen: But Lukeprog's posts are obviously brilliant, his output is great, even very focused readers like Chalmers find Luke to be very bright.

Which doesn't tell much about what they would have been were he under a more stringent diet. Another reasonable suspicion is that he was not actually modelling himself correctly, since he obviously does have an information diet

The Information Diet Challenge is to set yourself an information diet, explicitly, and follow it for a week.  

Many ways of countering biases have been proposed here, but I haven't found a post dealing with this specific, very low hanging fruit one. 

If you want inspiration, Ferriss has some advice here.

... but that is not the Positive Information Diet yet...

Information diets are supposed to constrain not everything you intake, but only what you intake instrumentally. If you just love reading about tensors and fairy tales, don't include them in what you won't avoid. What matters is to know that you'll avoid trying to learn programming by reading a programmer's tweet feed, avoid becoming a top researcher in psychology by reading popular magazines on it, and avoid reading random feeds on Facebook that don't relate to your goals in appropriate ways.

General form: I will Avoid spending my time reading/commenting things of kind (A)(Avoid), because I know that to reach my set of goals (G), the most productive learning time is doing (P) (Positve/Productive). 

So here is an attempt:

(G): Interact fruitfully with people at Oxford

(A): Facebook feeds that are not by them; News of any kind; Emails I can Postpone; Gossip; Books/articles not on Evolution of Morals, enhancement, AI; Wikidrifting; Family meal small talk; SMBC; 9gag; Tropes .... and a bunch of other stuff I don't have time or patience to list.

(P): Google scholar on the intersection between my research topic and theirs. Reading their papers by day, watching their videos by night. Re-read what I might help them with that was read before, list topics per person, write what to say about each topic.

What is wrong with this attempt is that (A) ends up being a negative list. A list of what what I do not want to intake. Since possibilities are infinite, this will give me ridiculous cognitive load, and that is a problem. So here is simple solution, which I used for a food diet before, and worked great:  Name not what you cannot do, but what you are allowed to do. Way fewer bits, way easier to check! 

Food example: I'll eat only plants, lean fish and chicken, nuts, fruits, whole pasta, beans and Chai Lattes.

We are better at checking for category inclusion than exclusion. There are so many available categories to exclude from that we don't feel bad that we "forgot" to check for that one. Then after you let yourself indulge in a tiny one, a small one doesn't seem that bad, and snowball effect does the rest. We sneak in connotations to make categories smaller, so our actions stay safely outside the scope of prohibition. Theoretically, we could do the reverse, but it is psychologically much harder. Just try to convince yourself that beef is "lean chicken" to see it.

So let us forget completely about (A). There is no kind or class of kinds to avoid. there is only G and P, and now there is also T, the time during which P is in force, since escape valves might be necessary to avoid "screw that" all-or-nothing effects.

An Improved attempt:

G: Interact fruitfully with people in Oxford

P: Google scholar on the intersection between my research topic and theirs. Reading their papers by day, watching their videos by night. Re-read what I might help them with that was read before, list topics per person, write what to say about each topic. Only Facebook them. 

T: 02:00-23:59 daily.

This is only for "computer use", where I'm most likely to do the wrong thing.

Now there is a simple to check list of things I want to do, I could be doing, and I'll try to do until G arrives. I can only do those. If x doesn't belong, don't do it, that simple. I'm free from midnight to two to do whatever, thus I don't feel enslaved by my past self.  No heavy cognitive load is burning my willpower candle (Shawn Achor 2010) by trying set theory gimmicks to get me to do the wrong thing. 

So please, take the:

          Positive Information Diet Challenge

Write your G's (goals) P's (positives) and T's (times), and forget about your A's (Avoids)