Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Rational Feed: Last Week's Community Articles and Some Recommended Posts

2 Post author: deluks917 02 October 2017 01:49PM

===Highly Recommended Articles:

Slack by Zvi Moshowitz - You need slack in your life. Slack lets you explore and invest. If you don't have slack you can't relax or uphold your morals. Fight hard to maintain your slack and don't let people or things take it away. Maya Millennial's lack of slack.

Personal Thoughts On Careers In Ai Policy And by carrickflynn (EA forum) - 3600 words. AI strategy is bottlenecked by hard research problems. Hence most people will find it hard to contribute effectively, even if they are very talented. Solving these problems has extremely high value. We should prepare to mobilize more talent once the blocking issues are solved. Operations work is still in high demand.

End Factory Farming by 80,000 Hours - Three hour podcast. How young people can set themselves up to contribute to scientific research into meat alternatives. Genetic manipulation of chickens. Skepticism of vegan advocacy. Grants to China, India and South America. Insect farming. Pessimism about legal or electoral solutions. Which species to focus on. Fish and crustacean consciousness.


Against Individual Iq Worries by Scott Alexander - "IQ is very useful and powerful for research purposes. It’s not nearly as interesting for you personally." IQ measurement problems. Even accurately measured IQ isn't that predictive.

Links: Hurly Burly by Scott Alexander - SSC links post. Copyright, genetic engineering, Autism, Machine Learning, Putin's fears of AI risk, the lesswrong relaunch and more


Dojo Bad Day Contingency Plan by Elo - Eleizer's discussion of why rationality theory isn't enough, you need to practice. An exercise about improving your mental on bad days.

Also Against Individual IQ Worries by Scott Aaronson - IQ tests tend to ask unclear questions and require you to reverse engineer what the test maker meant. Scott's own IQ was once measured at 106.

Predictive Processing by Entirely Useless - Responses to quotes from Surfing Uncertainty and Scott's review. A large focus is the "darkened room" problem.

Prosocial Manipulation by Katja Grace - Being calculating and guarded in communication is commonly considered manipulative and selfish. But many people's goals are pro-social, why do we assume manipulation is anti-social?

Humans As Leaky Systems by mindlevelup - "Fairly obvious stuff that probably lots of people are thinking about, but now put into simpler words (maybe). Basically, the idea that humans are affected by both ideas and the environment, and this is an important consideration in several models."

Dealism Futarchy And Hypocrisy by Robin Hanson - Policy conversations don't have to be about morality or terminal values. We can instead use tools like economics as a way to help people get whatever it is they want. We can push closer to the Pareto optimal frontier.

Debunking Iq Denial Ism by Grey Enlightenment - Criticisms of Scott's article on individual iq. People can change their socioeconomic status not their iq, IQ is more predicative than socioeconomic status, Feynman, Job titles are non-specific, low-iq 'computer' professions might be doing data entry. EQ isn't intrinsic and doesn't compete with IQ.

Harnessing Polarization by Robin Hanson - Capitalism channels status competition into productive enterprise. How can we similarly channel partisanship? Contests? Decision Markets?

Common Sense Eats Common Talk by Stefano Zorzi (ribbonfarm) - Missing the housing bubble. Falling for conformity. Seeing through invisible clothes. Advice: Test macro assumptions, beware of jargon, assume propositions that contradict common sense are wrong. Common talk and common sense and their failings.

Sabbath Hard And Go Home by Ben Hoffman - The Sabbth as easymode leisure. Unplugging while camping or on a meditation retreat feels natural. What is leisure? If you are unable to keep a Sabbath things are not ok, there isn't enough slack in the system.

Cognitive Empathy And Emotional Labor by Gordon (Map and Territory) - Affective empathy contrasted with Cognitive empathy. Cognitive empathy enables real emotional labor.

City Travel Scaling by Robin Hanson - Review of Geoffrey West's 'Scale'. Most visits to a location are from infrequent visitors who live nearby. Fractal piping systems have an overhead that only grows logarithmically with the size of the city. Evolution never found such efficient heating/cooling systems.

Travel Journal Hawaii by Jacob Falkovich - The Hawaiian language only has 40 syllables. Sales tax. Circadian Rhythm. Colonialism. The Hawaiian caste system. The best meal in the world. Don't quit your job to sell lemonade. Minimum wage ruined the pineapple industry.

Why I Quit Social Media by Sarah Constantin - Becoming stronger and less emotional since we live in a finite world with constrained resources. Social media: "It distances you from reality, makes you focus on a shadow-world of opinions about opinions about opinions; it makes you more impulsive and emotionally unstable; it incentivizes derailing conversations to fish for ego-strokes."


An Outside View Of Ai Control by Robin Hanson - Non-singularity scenarios where software performs almost all jobs. Software usually reflects the social organization of those who made it. Entrench designs and systems. Don't work on the control problem until its time. Human control and AI control. Most AI failures in this scenario will cause limited damage and can be handled after they occur.

Nonlinear Computation In Linear Networks by Open Ai - Floating point arithmatic is fundamentally non-linear near the limit of machine precision. OpenAI managed to exploit these non-linear effects with an evolutionary algorithm to achieve much better performance than a normal deep normal network on MNIST.

September 2017 Newsletter by The MIRI Blog - New MIRI paper on Incorrigibility and shitting off AI. Best posts from the intelligent agents forum. Links to videos and podcasts. MIRI personel updates and career opportunities in aI safety.

NBER Conference Artificial Intelligence by Marginal Revolution - Links to the program and videos. Tyler was there to comment on Korinek and Stiglitz.


What Happens To Cows In The Us by Eukaryote - "There are 92,000,000 cattle in the USA. Where do they come from, what are they used for, and what are their ultimate fates?"

Interim Update On Givewells Money Moved And Web Traffic In 2016 by The GiveWell Blog - Summary of influence, total money moved, money moved by charity.

Guardedness In Ea by Jeff Kaufman - As people and organizations gain prestige their communication becomes less open and more careful. Jeff has seen this happen in the EA community and dislikes the effects. However Jeff doesn't see a great alternative.

Trial Postponed by GiveDirectly - Give directly Kenya trial postponed due to political events.

===Politics and Economics:

Why White Identity Doesn't Work by Grey Enlightenment - Who counts as white. Race is secondary to ancestry and culture. No unifying cause or struggle. Whites may be biologically individualist. Too much infighting.

Comment on Oppressed Groups and Slack by Benquo - People who are oppressed often lack the slack to maintain their morals. Seven Samurai. This has the troubling implication that while we should listen to the oppressed the relatively privileged should maintain leadership. However it also implies that oppressed group's behavior will improve after enough time without a boot on their neck.

The OpenPhil Report On Incarceration by The Unit of Caring - "Our prison system isn’t just not-rehabilitative; it is anti-rehabilitative. It traumatizes and retraumatizes people and severs their connections to people and opportunities within the law and abuses them and breaks social trust and produces crime which is then used to justify longer prison sentences which produce more crime."

On The Fetishization Of Money In Galts Gulch by Ben Hoffman - Danny Taggart and Galt feel they can't ethically become lovers until they rectify a power imbalance. Danny solves this problem by becoming Galt's house-maker and cook. Most people's intuition is that employment creates a power imbalance, it doesn't solve one. What is going on?

Seasteading 2 by Bayesian Investor - "The book’s style is too much like a newspaper. Rather than focus on the main advantages of seasteading, it focuses on the concerns of the average person, and on how seasteading might affect them. It quotes interesting people extensively, while being vague about whether the authors are just reporting that those people have ideas, or whether the authors have checked that the ideas are correct. Many of the ideas seem rather fishy."

What Is Going On With The Alt Right by Grey Enlightenment - Reasons the alt-right is falling apart: Trump back-peddling or softening on campaign promises, The civil war between the-lite, alt-medium, and alt-right, Slow news cycle and brevity of ideas, Botched rallies and poor branding, The alt-right losing its official Reddit sub, the right is more intellectually diverse than the left.

Milgram Replicates by Bryan Caplan - Milgram's shock study replicated well in 2009. Since 79% of people who pushed past the subjects first verbal protest went to the end of the range the replication stopped earlier than Milgram.


Summary Of Reading July September 2017 by Eli Bendersky - Book reviews: Stats, genetics, Winnie the Pooh, Zen and other topics.


Creating Trump by The Ezra Klein Show - "How the Republican Party created Trump, how Trump won, and what comes next. As Dionne says in this interview, the American system was "not supposed to produce a president like this,” and so a lot of our conversation is about how the guardrails failed and whether they can be rebuilt."

Rs 194 Robert Wright On Why Buddhism Is True by Rationally Speaking - "Why Buddhism was right about human nature: its diagnosis that the our suffering is mainly due to a failure to see reality clearly, and its prescription that meditation can help us see more clearly. Robert and Julia discuss whether it's suspicious that a religion turned out to be "right" about human nature, what it means for emotions to be true or false, and whether there are downsides to enlightenment."

Robert Wright by EconTalk - "The psychotherapeutic insights of Buddhism and the benefits of meditation and mindfulness. Wright argues our evolutionary past has endowed us with a mind that can be ill-suited to the stress of the present. He argues that meditation and the non-religious aspects of Buddhism can reduce suffering and are consistent with recent psychological research."

Burning Man by The Bayesian Conspiracy - How much does burning man live up to its principles, changes over time, finding out you aren't gay in your twenties, marriage. Burning Man advice: Go with a camp you like, don't have plans just wander around and get involved in whats interesting.

The Fate Of Liberalism by Waking Up with Sam Harris - "Mark Lilla about the fate of political liberalism in the United States, the emergence of a new identity politics, the role of class in American society"

Comments (13)

Comment author: gjm 03 October 2017 11:53:24AM 4 points [-]

The "Grey Enlightenment" post about Scott's "Against individual IQ worries" seems to me to have a big mistake at its heart. It quotes Scott's comparison between "my IQ isn't very high" and "my family isn't rich" (both of which are disadvantages in life, neither of which is insurmountable for most purposes) and says

The mistake he is he is conflating IQ (which is biological and thus intrinsic to the individual) with parental wealth (which is not). [...] One can change their socioeconomic standing; one cannot change their IQ.

which misses the point completely. You can change your socioeconomic standing, but you can't change your parents' wealth. This is exactly parallel to how you can change (e.g.) how well you understand physics, or how many of Shakespeare's plays you've read, or how effective a debater you are, even though you can't change your IQ and IQ makes a difference to your ability to learn physics or Shakespeare or debating.

(I am using "IQ" as a shorthand for "overall brainpower as measured by things like IQ tests" and assuming arguendo that it is in fact true that you have negligible ability to change this other than perhaps by improving a specific IQ-test-taking ability that doesn't transfer to other intellectual tasks.)

I guess the Greyly Enlightened One might want to say: "Sure, you can improve those things, but if your IQ isn't high enough then some things are completely out of reach to you, whereas bad family circumstances don't do that. If your IQ is 110 then you are never, ever going to understand string theory." But it seems to me that (1) there are actually very few things that are completely out of reach to people of good but not outstanding intelligence, which I take it is roughly the group Scott's mostly addressing, and (2) the information most people have and might possibly worry about is not really of the form "my coefficient of general intelligence is such-and-such" but rather "I took an IQ test once and the result was such-and-such". To the latter I reply: Richard Feynman, 124. Scott Aaronson, 106. Yes, those results don't give an adequate description of Feynman's or Aaronson's brainpower. That's the point.

Comment author: IlyaShpitser 03 October 2017 12:19:09PM *  2 points [-]

"If your IQ is 110 then you are never, ever going to understand string theory."

I always wondered about these types of claims. String theory is just math applied in a particular way. So we can try to figure out exactly where the line is:

Can you understand calculus with "IQ 110" (I think clearly so)?

How about analysis of the complex plane?

How about linear algebra, hermitian matrices, etc?

How about group theory?

String theory stuff is very complicated, but it's just made up of this type of math put together in a particular way. Some of it just takes time to internalize, but there is never any "magic sauce" about any of the specific parts, I don't think.

One might say smart folks take less long to internalize, but my experience has been truly internalizing complex math is a bit of a slow process for everyone.

I think if Scott Aaronson once took an IQ test and got 106, that should tell you everything you need to know about how good this proxy is for "complex cognition stuff."

I think folks in the rationality-sphere have a Mensa-like obsession with this number. Folks in Mensa do daily puzzles and worry about their IQ, folks in academia/industry publish papers and contribute to the intellectual conversation or create things and contribute to civilization. I humbly submit the latter is a better use of time.

Plus, a resume is a much better proxy for intelligence than IQ -- more bits.

Comment author: gjm 03 October 2017 02:03:22PM 1 point [-]

Yeah, I'm also by no means certain that hypothetical-Greyly-Enlightened-One would be right in claiming that an IQ of 110 means you'll never understand string theory. I think it's somewhat plausible, though; it feels to me as if there are things in mathematics that I understand quite easily, and things I can get my head around with substantial effort, and things that are just beyond me, and it feels as if the problem with the latter really is a straightforward lack of brainpower. Of course this is all just introspection, and we know how unreliable that is, and my guess is that at least some of the things I think are "just beyond me" I could in fact get my head around if I put in a heroic effort. But, still, it seems pretty damn plausible that some things are just too hard for some people to be realistically able to grasp them, and that in so far as there is any such thing as "your IQ" it's a decent proxy for that.

(Of course there is such a thing as mathematical talent in particular areas, and there are blind spots in particular areas. I don't think even Grey Enlightenment Guy would actually deny the possibility that someone might have a lowish measured IQ but specific extraordinary abilities that might enable them to get the hang of string theory or whatever. But -- and this is not

I think it's worth separating the propositions (1) "whatever IQ tests are trying to measure, they don't measure it anything like perfectly" and (2) "whatever IQ tests are trying to measure, it isn't perfectly predictive of actual ability to do intellectually difficult things". The fact that Scott Aaronson once scored only 106 on an IQ test is strong evidence for some combination of those things but leaves it entirely open how much of it is measurement error and how much is weakness of even-ideally-measured IQ.

Strongly agree with the last section. I think a lot of interest in IQ arises from sources we should be skeptical of:

  • If you have reason to think your own IQ is high, thinking about IQ can feel good. (This is the main source of the Mensa problem, I take it.)
  • Reducing something complicated to a single numerical measure is appealing.
  • (I raise this one with great caution, but I do think it is part of the picture in some cases and I think Mr Grey Enlightenment might be one of them.) There is some evidence for correlations between measured IQ and race, and for whatever reason some people are favourably disposed towards anything that bolsters "traditional" views about race.

Of course for each of these there's a corresponding bad reason for not being interested in IQ. If you've had your IQ measured and it didn't come out very high, you may be motivated to downplay the importance of IQ; just as some people love to simplify and quantify, some people love to complicate and obfuscate; and if you have heard that measured IQ seems to vary with race and dislike racial prejudice, you may want to avoid thinking about IQ in case it leads to some sort of moral corruption. My guess is that the first two of these things, at any rate, are more likely to lead to overemphasis than underemphasis on IQ in the rationalsphere. (I bet the third goes both ways but is more likely to lead to underemphasis.)

Comment author: IlyaShpitser 03 October 2017 02:21:00PM *  1 point [-]

The thing about race and intelligence (aside form the fact that it's a hopelessly toxic topic) is that most folks making claims relating the two can't possibly have the data to be confident about anything. Intelligence is complicated, very complicated. "Race" is complicated, too. I don't have to obfuscate anything here, because genetics is inherently weird and messy, and so are brains.

So if folks sound confident, or make strong claims they are either confused or racist or both. The everburning tire fire of slatestar's comment section (when it comes to this topic) is a prime example of what I am talking about.

Comment author: gjm 03 October 2017 02:54:34PM 1 point [-]

I am inclined to agree, except that rather than "confused or racist" I would say "confused or strongly influenced by prior prejudices", and note that one can have prior prejudices against racial IQ differences as well as for them. That is, if someone declares confidently that of course there are no differences between races in mental qualities, that's evidence of bad thinking for all the same reasons as if they declare confidently that there are such differences.

A prior prejudice against racial brainpower differences is much nicer than a prior prejudice in favour of such differences, and probably people with the former kind of prejudice are mostly better people than ones with the latter kind, but letting either induce confident strong opinions about race-and-intelligence is the same kind of mistake.

Comment author: Lumifer 03 October 2017 03:43:32PM 1 point [-]

A prior prejudice against racial brainpower differences is much nicer than a prior prejudice in favour of such differences

Why so? Certainly one is much more socially acceptable than the other, but that's not the usual definition of "nice".

Is a "prior prejudice" against, say, sex differences in upper body strength "much nicer" than a prior prejudice in favour?

Comment author: Lumifer 03 October 2017 03:40:22PM 0 points [-]

most folks making claims relating the two can't possibly have the data ... So if folks sound confident, or make strong claims they are either confused or racist or both

There is a big difference between "most folks" and "all folks".

Just because e.g. Twitter is an permanent garbage fire about most complicated topics does not mean that everyone with an opinion about one of these topics is confused (or racist).

Comment author: gjm 03 October 2017 03:45:34PM 0 points [-]

I'm puzzled by this comment. Ilya didn't say anything about "e.g. Twitter" (he did mention the Slate Star Codex comments section, which is a far cry from Twitter), and he didn't say anything about "everyone with an opinion".

Comment author: Lumifer 03 October 2017 03:50:30PM 0 points [-]

I am reading "So if folks sound confident" as "So if [all/any] folks sound confident". I believe Ilya's point is that no one should have strong opinions on race/IQ and I disagree with it.

Comment author: Lumifer 03 October 2017 03:36:25PM *  0 points [-]

it seems pretty damn plausible that some things are just too hard for some people to be realistically able to grasp them


a lot of interest in IQ arises from sources we should be skeptical of

Sure, those are good reasons to suspect motivations, but not good reasons to stay away from IQ issues completely.

Ignoring IQ differences in social sciences (including economics) is likely to lead to... suboptimal conclusions and recommendations.

Crudely speaking, dumb people exist in large numbers and one should not pretend that this is not so.

Comment author: Allyne43 09 October 2017 05:05:20PM 0 points [-]

Sometimes i just don't know what to say actually, it's important to know this but at the same time i don't have enough strength to read it anymore... so many things that makes me feel so pissed off!

Comment author: WalterL 02 October 2017 05:28:44PM 0 points [-]

The article about Slack is really good, thanks for linking.

Comment author: gjm 03 October 2017 11:38:48AM 1 point [-]

FYI, it's on the LW2 site as well as Zvi's blog. So are some of the other things here.