I seem to have high karma, but don't know why. Looking through my contribution history, I seem to only have a total of 47 net upvotes on anything I've ever posted, but have 74 karma points, including 10 in the last 30 days. Looking at the LW wiki FAQ, it says that you can get 10 karma per upvote if you post in main, but I haven't done that. Does anyone know why this might be happening?

I don't have any sensible way of learning about current affairs. I don't consume broadcast or print news. Most news stories reach me through social media, blogs, word of mouth or personal research, and I will independently follow up on the ones I think are worthy of interest. This is nowhere near optimal. It means I will probably find out about innovations in robotic bees before I find out about natural disasters or significant events in world politics.

Regular news outlets seem to be messy, noisy attention traps, rather than the austere factual repositories I wish them to be. Quite importantly, there seems to be a lot of stuff in the news that isn't actually news. I'm pretty sure smart people with different values will converge on what a lot of this stuff is.

Has this problem been solved already? I'm willing to put in time/effort/money for minimalist, noise-free, sensibly-prioritised news digest that I care about.

ETA: Although I haven't replied to all these responses individually, they seem very useful and I will be following them up. Thanks!

Okay, I know next to nothing about Haskell, and next to nothing about provability logic, so maybe what I'm about to ask doesn't make any sense, but here's something that's making me very curious right now. How do I implement a function like this:

loeb :: Prov p => p (p a -> a) -> p a
loeb = ???

using some typeclass like this:

class Prov p where
dup :: p a -> p (p a)
mp :: p (a -> b) -> p a -> p b
???

The idea is that the implementation of loeb should follow the steps of this proof, and the methods of Prov should correspond to the assumptions of that proof. The question was inspired by this post by sigfpe, but he thought that the class should be Functor, which seems wrong to me.

Apologies for the formatting, it turns out LW collapses whitespace even in preformatted blocks.

How do I build the habit of writing down a fleeting thought that seems interesting? Way too often I notice that I just wanted to do something or write something down. Or should I just accept the thought as gone?

Here are two bookmarklets that have really helped my article-reading workflow. I named the bookmark for #1 "Clean" and #2 "Squirt":

1. javascript:window.location.replace("<http://justread.mpgarate.com/read?url=>" + escape(document.URL))
2. javascript:(function()%7Bsq%3Dwindow.sq%3Dwindow.sq%7C%7C%7B%7D%3Bif(sq.script)%7Bsq.again()%3B%7Delse%7Bsq.bookmarkletVersion%3D%270.3.0%27%3Bsq.iframeQueryParams%3D%7Bhost:%27//squirt.io%27,userId:%27f98bd133-7ae4-45da-be5d-16a4919b902c%27,%7D%3Bsq.script%3Ddocument.createElement(%27script%27)%3Bsq.script.src%3Dsq.iframeQueryParams.host%2B%27/bookmarklet/frame.outer.js%27%3Bdocument.body.appendChild(sq.script)%3B%7D%7D)()%3B

#1 is a nice, simple frontend for the Readability API. Just enter the URL of a page with something you want to read on it, and it extracts the content without any sidebars, ads, or other junk and gives it to you in an easy-to-read format.

#2 is Squirt, a speed-reading application that takes the text of any webpage and displays it to you one word at an adjustable speed. The default is 450wpm I think, but after you make an adjustment, it remembers what speed you want for next time. If you need to read a part more carefully or go back because you missed something, that's easy. Hit the spacebar to pause and it will show the context, then use the left and right arrows to move around. Hit the spacebar again to resume speed reading. Another awesome feature is that it tells you exactly how long it will take to finish if you don't stop, so you can decide if it's worth your time or not.

The two work really well together, as squirt alone will sometimes grab text you don't want. What I do is "Clean" a page by clicking on the bookmarklet, and then sometimes hit "Squirt" to speed read it.

Try it out and let me know what you think!

I like the notion of the Superintelligence reading group: http://lesswrong.com/lw/kw4/superintelligence_reading_group/. But the topic of AI doesn't really interest me much.

A reading group on some other topic that is more along CFAR's lines than MIRI's would. For example, reading recent studies of cognitive bias would be interesting to me. Discussion on how practically to combat them might evolve from discussing the studies.

Max L.

If you liked Scott Alexander's essay, Meditations on Moloch, you might like this typographic poster-meme I made. It was a minor success on Facebook.

(If you haven't read Scott Alexander's essay, Meditations on Moloch, then you might want to check it out. As Stuart Armstrong said, it's a beautiful, disturbing, poetical look at the future.)

Since 23andme has been prohibited from giving health-related genetic reports, is there anyone else (outside the FDA's jurisdiction) who provides similar services?

Edit: I have found Promethease, which works with 23andme's raw data. I'm still interested in additional options.

Edit2: This page lists various 23andme competitors, although it was last updated in early 2013. More recent information is appreciated.

A medical issue is a problem if the patient recognises it as one. If a patient suffers from something that is not recognised as medical problem we call it hypochondria. Is there the concept of something we see as medical problem but the patient does not realise as one e.g. because they don't know that their condition is not normal?

We managed to reduce performance on any number of tests to essentially a single number, g, together with a couple more for domain-specific skill. We managed to reduce the huge variation in personalities to five numbers, the OCEAN dimensions. I even recall reading that there is quite some correlation between those five numbers and that they might be reduced to a single one but I can't find the source any more.

Can we construct a whole host of other, similar numbers, like "math skills" and thus measure the impact of education and aging?

Another number I have in mind is, can we construct three numbers general health gh, mental health mh and physical health ph, and measure their correlations? I have the vague observation that medical issues tend to cluster, that is people with mental issues tend to not only exhibit any one of ADHD, depression, OCD and so on, but more than one of them. Similarly I have the impression that people tend to complain of many physical symptoms at once.

I seem to recall that BMI and/or WHR tend to be excellent predictors of physical health. Together with a couple of more measures these predictions can further be improved. The advantage of having a single number would be for research purposes on population health and it is easier to have a single mumber for personal assesment.

An interesting paper. The abstract says:

Rationality leads people to imitate those with similar tastes but different information. But people who imitate common sources develop correlated beliefs, and rationality demands that later social learners take this correlation into account. This implies severe limits to rational imitation. We show that (i) in most natural observation structures besides the canonical single-file case, full rationality dictates that people must “anti-imitate” some of those they observe; and (ii) in every observation structure full rationality dictates that people imitate, on net, at most one person and are imitated by, on net, at most one person, over any set of interconnected players. We also show that in a very broad class of settings, any learning rule in which people regularly do imitate more than one person without anti-imitating others will lead to a positive probability of people converging to confident and wrong long-run beliefs.

It Ain't Necessarily So: Why Much of the Medical Literature Is Wrong

Some of the material will be familiar, but there are examples I hadn't seen before of how really hard it is to be sure you've asked the right question and squeezed out the sources of error in the answer.

What follows is what I consider to be a good parts summary-- if you want more theory, you should read the article.

Consider a study published in the NEJM that showed an association between diabetes and pancreatic cancer.[3] The casual reader might conclude that diabetes causes pancreatic cancer. However, further analysis showed that much of the diabetes was of recent onset. The pancreatic cancer preceded the diabetes, and the cancer subsequently destroyed the insulin-producing islet cells of the pancreas. Therefore, this was not a case of diabetes causing pancreatic cancer but of pancreatic cancer causing the diabetes.

....

To illustrate the point, consider the ISIS-2 trial,[8] which showed reduced mortality in patients given aspirin after myocardial infarction. However, subgroup analyses identified some patients who did not benefit: those born under the astrological signs of Gemini and Libra; patients born under other zodiac signs derived a clear benefit with a P value < .00001.

I guessed at a seasonal effect, but Gemini and Libra aren't adjacent signs.

The frequency of these false-positive studies in the published literature can be estimated to some degree.[2] Consider a situation in which 10% of all hypotheses are actually true. Now consider that most studies have a type 1 error rate (the probability of claiming an association when none exists [ie, a false positive]) of 5% and a type 2 error rate (the probability of claiming there is no association when one actually exists [ie, a false negative)] of 20%, which are the standard error rates presumed by most clinical trials. This allows us to create the following 2x2 table.

I didn't realize that the false negative effect (not seeing a relationship when there actually is one) is higher than the false positive rate. This might mean that a lot of useful medical tools get eliminated before they'can be explored.

Also (credit given to Seth Roberts), if a minority of people respond very well to a treatment being tested, this is very unlikely to be explored because the experiment is structured to see whether the treatment is good for people in general (actually, people in general in the group being tested). This wasn't in the NEJM piece.

One classic example of selection bias occurred in 1981 with a NEJM study showing an association between coffee consumption and pancreatic cancer.[15] The selection bias occurred when the controls were recruited for the study. The control group had a high incidence of peptic ulcer disease, and so as not to worsen their symptoms, they drank little coffee. Thus, the association between coffee and cancer was artificially created because the control group was fundamentally different from the general population in terms of their coffee consumption. When the study was repeated with proper controls, no effect was seen.[16]

....

Information bias, as opposed to selection bias, occurs when there is a systematic error in how the data are collected or measured. Misclassification bias occurs when the measurement of an exposure or outcome is imperfect; for example, smokers who identify themselves as nonsmokers to investigators or individuals who systematically underreport their weight or overreport their height.[17] A special situation, known as recall bias, occurs when subjects with a disease are more likely to remember the exposure under investigation than controls. In the INTERPHONE study, which was designed to investigate the association between cell phones and brain tumors, a spot-check of mobile phone records for cases and controls showed that random recall errors were large for both groups with an overestimation among cases for more distant time periods.[18] Such differential recall could induce an association between cell phones and brain tumors even if none actually exists.

::::

An interesting type of information bias is the ecological fallacy. The ecological fallacy is the mistaken belief that population-level exposures can be used to draw conclusions about individual patient risks.[4] A recent example of the ecological fallacy, was a tongue-in-cheek NEJM study by Messerli[19} showing that countries with high chocolate consumption won more Nobel prizes. The problem with country-level data is that countries don't eat chocolate, and countries don't win Nobel prizes. People eat chocolate, and people win Nobel prizes. This study, while amusing to read, did not establish the fundamental point that the individuals who won the Nobel prizes were the ones actually eating the chocolate.[20]

On the other hand, if you want to improve the odds of your children winning a Nobel, maybe you should move to a chocolate-eating country.

.....

A 1996 study sought to compare laparoscopic vs open appendectomy for appendicitis.[29] The study worked well during the day, but at night the presence of the attending surgeon was required for the laparoscopic cases but not the open cases. Consequently, the on-call residents, who didn't like calling in their attendings, adopted a practice of holding the translucent study envelopes up to the light to see if the person was randomly assigned to open or laparoscopic surgery. When they found an envelope that allocated a patient to the open procedure (which would not require calling in the attending and would therefore save time), they opened that envelope and left the remaining laparoscopic envelopes for the following morning. Because cases operated on at night were presumably sicker than those that could wait until morning, the actions of the on-call team biased the results. Sicker cases preferentially got open surgery, making the outcomes of the open procedure look worse than they actually were.[30] So, though randomized trials are often thought of as the solution to confounding, if randomization is not handled properly, confounding can still occur. In this case, an opaque envelope would have solved the problem.

Remembering that humans aren't especially compliant is hard.

From reading Guinea Pig Zero: The Journal for Human Research Subjects-- human beings are not necessarily going to comply with onerous food regimes. I expect that most who don't simply don't want to, but the magazine had the argument of not wanting to comply because the someone who's a human research subject is never going to be able to afford treatment based on the results of the research.

I have a notion that an FAI will be able to create better friends and lovers for you than actual humans could be. Family would be a more complex case if you value the history as well as the current experience.

I'm not talking about catgirls-- if some difficulties in relationships are part of making relationships better in the long haul, then the FAI will supply difficulties.

If people eventually have relationships with FAI-created humans rather than humans generated by other means, is this a problem?

I have yet to find any thoughts on Effective Altruism that do not assume vast amounts of disposable income on the part of the reader. What I am currently trying to determine are things like 'at what point does it make sense to give away some of your income versus the utility of having decent quality of life yourself and insuring against the risk that you end up consuming charitable resources because something happened and you didn't have an emergency fund'. Does anyone know of any posts or similar that tackle the effective utilitarian use of resources when you don't have a lot of resources to begin with?

It seems like there are a lot of fan-fiction fans here. Fan-fiction fans, I am curious as to what draws you to the fan-fiction of which you are fans. Is part of it that you're fans of other fan-fiction fans? I guess that depending on the cosplay you could even be a fan of fan-fiction fans' fans.

I remember seeing this organization on LW but cannot find it again or remember the name: it was a for-profit school-like entity that does a short training program (might have been 6 six week, maybe 3 months, that range), which is free upfront and takes their payment entirely as a percentage of the salary from the job they place you in afterward. If I remember correctly, it is run in the Bay Area and took a small pool each session, with a school-like application process.

Can anyone point me to this?

I'm posting here on behalf of Brent Dill, known here and elsewhere as ialdabaoth-- you may have enjoyed some of his posts. If you read the comments at SSC, you'll recognize him as a contributor of rare honesty and insight. If you'd had the chance to talk with him as much as I have, you'd know he's an awesome guy: clever, resourceful, incisive and deeply moral. Many of you see him as admirable, most as relatable, some as a friend, and more, I hope, as a member of our community.

He could use some help.

Until last Thursday he was gainfully employed as a web developer for a community college in Idaho. Recently, he voluntarily mentioned to his boss that he was concerned that seasonal affective disorder was harming his job performance, who mentioned it to his boss, who suggested in all good faith that Brent should talk to HR to see if they might help through their Employee Assistance Program. In Brent's words: "Instead, HR asked me a lot of pointed questions about when my performance could turn around and whether I wanted to work there, demanded that I come up with all the solutions (after I admitted that I was already out of brainpower and feeling intimidated), and then directed me to turn in my keys and go home, and that HR would call me on Monday to tell me the status of my employment." Now, at the end of the day Tuesday, they still haven't let him know what's happening, but it doesn't look good.

I think we can agree that this is some of the worst horseshit.

On the other hand, he's been wanting to get out of Idaho and into a city with an active rationalist community for a while, so in a sense this is an opportunity. Ways to help: Brent needs, in order of priority: a job, a place to stay, and funds to cover living and moving expenses-- details below. Signal boosts and messages of support are also helpful and appreciated. Ways NOT to help: Patronizing advice/other-optimizing (useful information is of course welcome), variations on 'cool story bro' (the facts here have been corroborated to my satisfaction with hard-to-fake evidence), disrespect in general.

1. Job: Leads and connections would help more than anything else. He's looking to end up, again, in a good-sized city with an active rationalist community. Candidates include the Bay Area, New York, Boston, Columbus, San Diego, maybe DC or Ann Arbor. He has an excessively complete resume here, but, in short: C#/.NET and SQL developer, also computer game development experience, tabletop board/card game design experience, graphic art and user interface experience, and some team leadership / management experience.

2. Crash space: If you are in one of the above cities, do you have/know of a place for a guy and his cat? How much will it cost, and when will it be available? Probably he'll ultimately want a roommate situation, but if you're willing to put him up for a short time that's also useful information.

3. Funds: Brent is not now in immediate danger of going hungry or homeless, but a couple of months will exhaust his savings, and (although it is hard to know in the current state of things) he has been told that the circumstances constitute "cause" sufficient to keep him from drawing unemployment. Moving will almost certainly cost more than he has on hand. There is a possible future in which he runs out of money stranded in Idaho, which would be not good.

If you feel moved to help, he has set up a gofundme account here. (The goal amount is set at his calculated maximum expenses, but any amount at all would help and be greatly appreciated-- he would have preferred not to set a funding goal at all.) Though Brent has pledged to eventually donate double the amount he raises to Effective Altruist causes, we wouldn't like you to confuse contributing here with charitable giving. Rather, you might want to give in order to show your appreciation for his writing, or to express your solidarity in the struggles and stigma around mental illness, or as a gesture of friendship and community, or just to purchase fuzzies. Also, you can make him do stuff on Youtube, you know, if you want.

Thank you so much for your time and kindness. -Elissa Fleming

According to the efficient market hypothesis index funds should be the best way for the average person to gain a return from investment. Now there is a plethora of indices to invest in. How should one find the 'best' one?

Further, only a relatively small part of return generating assets are captured in publically tradeable assets. What about private equity and real estate, huge parts of the economy?