The benefits of this realization have spanned many aspects of my life, and have substantially increased my altruistic human capital.
This is really interesting. Can you give some concrete examples from different aspects and explain a bit how your capital has increased?
Yes, I agree that if it’s a substantially better predictive model it would be very useful. Do you have any concrete examples of, say, beliefs you came to using many many weak arguments that you would not have come to using a single strong argument, which later turned out to be true? Or ways in which you started modeling people differently and how specifically those improved models were useful? (My prior for “one thinks one is improving instrumental rationality ⇒ one is actually improving instrumental rationality” is rather low.)
Brienne, my roommate and I would like to do this, but our schedule won’t match yours because we’re time constrained. (We already started taking baseline data, but we’re doing the adaptation this weekend as opposed to on Monday.) Should we still fill out your form or will the timing mess things up?
Also, my PMV score gets about ~50ms lower when I use a mouse as opposed to a trackpad. Hopefully this isn’t an issue, but people should make sure they’re using consistent hardware.
Worth adding that most people don’t actually need all of GTD, since the system was designed for a particular type of corporate drone. It’s really the attitude about doing things that’s important to take away. For instance, I don’t keep lists of projects or a “tickler” file—for me, the most important things were the weekly strategic reviews, “next actions” idea, and having a very small number of inboxes that I regularly clear out and categorize.
Are complex macro models how the world actually works? What are our most successful macro models, and how successful have they been? My impression (based off not that much data, admittedly) was that this applied to all macro, not just undergrad.
So I think the complaint about “clever accounting” is wrong. The counter-argument is somewhat technical, so it might still incur a PR penalty, but it’s not actually true.
As an individual donor considering where to donate based on ROI, you don’t actually care about the ROI that includes fundraising, because the fundraising doesn’t target you. The fundraising is targeted at people who currently don’t choose their charity based on ROI, so it actually makes sense to treat it as a completely exogenous process for the purposes of marginal ROI calculation.
(EDIT: because the average ROI is the integral of the marginal ROI, it’s clear that the marginal ROI has to get lower somewhere for the math to work out. I think the right way to do the accounting is to say that the low-marginal-ROI donations are those that come from the people whose minds are changed by Effective Fundraising. I smell some weird decision theory stuff going on behind the scenes here but I’m not quite sure if it’s interesting or important.)
A helpful thought experiment is replace “fundraising org” with “org that creates additional people who donate to the highest-ROI charity” (because that’s what they’ll be doing; if EF discovers a higher-ROI charity, then they’ll switch to fundraising for that one instead, hopefully). This might make it more clear that you shouldn’t include fundraising costs in the cost of a particular charity—the charities “just happen” to be where the newly-minted altruists are donating their money.
EDIT: By the way, this objection has been raised both times EF has been brought up to a wider audience. So if anyone can think of a cleaner/more intuitive way to explain the counter-argument above, it would probably be quite high-value.
I just realized that my PMV score also varies not just with mouse usage, but also with which OS I’m running (Windows gave me a ~40ms higher score). None of my data is contaminated, but just thought I’d warn you in case anyone at Leverage is dual booting.
This question seems to be distinct from some of the others in that it requires insight to answer, not just insight to ask.
When you ask something like “what skill am I actually training”, the answer is usually fairly obvious; the insight (that we often teach people based on some heuristic like “what would a teacher do” rather than the obviously correct “how can I get this person to reliably perform this skill”) is encapsulated in the question.
That said, I don’t typically ask myself this question. Is this a question you routinely ask yourself? If so, do you find that it’s actually helpful to ask it, as distinct from the benefits you would get from e.g. a general cue like “think harder about this problem”?
I’ve heard other people mention this approvingly, but I don’t know if the actually did it. It’s on my list of things to try at some point, though.
This is probably what you meant, but it might be more productive (and slightly less Hard Mode) to be able to choose “do immediately, do never, or break up into subtasks” while you’re actually going through the list.
Also, there are some tasks that I actually do want to postpone because they’re not my comparative advantage (for example, it’s raining right now, so I want to postpone errands until later tonight when it stops). That’s prone to rationalization, so I’m not sure how a system could handle that, but it’s something to think about.
What model of the world/way of thinking/worldview generated the things this person is saying?
What game am I playing? What are the rules (or, more relevantly, what aren’t) and how do I win?
What is the goal of this conversation? (Note: you must be careful to answer with the actual goal, not the goal you think you should have or the goal you would answer with in public. For instance, if your answer is just “truth seeking” you are wrong with very high probability.)
Maybe this wasn’t your intent, but framing this post as a rebuttal of Chapman doesn’t seem right to me. His main point isn’t “Bayesianism isn’t useful”—more like “the Less Wrong memeplex has an unjustified fetish for Bayes’ Rule” which still seems pretty true.
This is quite related to ignoring information that doesn’t fit into a framework, but another common logical bias is forcing information into your framework when it doesn’t fit.
My most obvious personal encounter with this was in my high school English classes, where my teacher frequently criticized me for having an “overbearing” interpretation of the text. For a perhaps more relatable example, I’ve known people who have just learned about status to interpret absolutely every behavior in the context of status, even when that doesn’t quite fit.
You may be interested in this Quora answer by Venkatesh Rao. An excerpt directly relevant to your question, with which I wholeheartedly agree:
When I read a [David Foster Wallace] passage, it is like looking at a pinprick-sharp photograph, compared to my own blurry photographs. He unerringly picks words to use that simply work 100x better than my choices. It’s like he has a 15 megapixel camera and a tripod, while I am using a 3 megapixel point-and-shoot. A bigger vocabulary isn’t enough. The skill lies in matching words to needs.
In fact his language is so precise that it makes his writing almost too rich to read. I’ve never finished any of his novels because they are too rich for me. My brain can’t handle it.
And this isn’t just at the word level. His sentences, paragraphs and chapters are massively precise as well. James Joyce is another example. His prose has been described as having the precision of poetry (an amazing feat, given that typical good poetry is generally 100x more precise than typical good prose, and Ulysees is HUGE).
Here’s a sample of David Foster Wallace and of Joyce. I’m surprised not to see at least the former mentioned here already.
Okay, so it seems pretty clear that he can’t do math (or lacks a good math editor). That’s disappointing, especially since he has a degree in it.
But I don’t think either of the other two speaks to lack of conceptual precision: the Language Hat post is essentially complaining that he’s a prescriptivist, which seems conceptually fine (if possibly wrong) to me. Meanwhile the eXiled one is an extensive (and in places unbelievably hatchet-y) attack on him for I’m not quite sure what—pretending to be folksy and chummy while secretly using big words?
(I admit that I skimmed because the author didn’t seem to have a point other than rambling about how the sponsor of the guy who made a movie of a book in the same genre as Infinite Jest once did some questionable political things. So maybe I missed something.)
Yes, and the thing that we care about also varies even within a single brain. For instance, my verbal loop/System 2 believes basically Parfit’s theory of identity, and given enough time to reflect I would probably make decisions based on it, but my System 1 is still uncomfortable with the idea.
That said, this is still worth arguing about. By analogy to ethics, the thing we (intuitively) care about there is some really complicated implicit definition, but that definition is inconsistent and probably leads to the ethical equivalent of dutch-booking, so many people here choose to overrule their intuitions and go with utilitarianism when the two conflict. There’s not necessarily any theory of ethics that’s definitively right, but there are certainly theories that are definitively wrong, and trying to construct a logical ethics that coincides as much as possible with our intuitive beliefs will help us iron out bugs and biases in those intuitive beliefs.
The same goes for identity—that’s why we should try to construct a theory that we think is consistent and also captures as much of our intuition as possible.
For the examples I can think of (mostly philosophy of mind), it seems to me that the sciences would have emerged whether or not any progress was made while it was still considered the domain of philosophy. Are there better examples, where the “philosophical” progress was actually important for the later “scientific” progress?
You missed one type of value, which is peer group. Most of the more difficult courses I’ve taken have given me at least some value by granting better access to more smart, competent people. This has a number of benefits:
Better human capital. By watching how more competent people work and think, you can often pick up useful study habits and better techniques for the subject you’re studying. I’ve found this especially true in CS classes, where I’ve had this experience from both sides, e.g. teaching classmates how to use Git and picking up C coding style and tricks from better programmers.
Advice/planning help: both more advanced students and instructors can be very useful for the academic advice they provide later. Knowing talented students has given me info about several excellent courses, as well as summer opportunities, I wouldn’t otherwise have known about. A professor who can become a good mentor is also invaluable (although if you don’t plan to go into academia, you may have trouble finding a professor who can sympathize with your goals enough to give you good advice).
Better networking later: this is hopefully obvious, but knowing smart, competent people is useful for pretty much everything ever.
Friends/fun: people who think fast and have good background are less predictable/more interesting to talk to.
This suggests the following advice:
Seek out classes that signal difficulty or are known to be challenging. Many colleges have an accelerated introductory math sequence that will be good for this; in computer science, courses on operating systems and compilers are often known for their difficulty and attract the best/most interested students.
Avoid courses that are hard requirements for anything. These courses are the worst at filtering for good students because they’re required. Additionally, they’ll probably be taught at a low enough level that the good students will be disengaged and hard to spot.
Among sets of required alternatives, look for smaller courses. The largest course is probably the lowest common denominator. This is particularly important for general-education courses, in my experience.
Data from my experience at Harvard: variability between instructors is very high (probably because many of them got professorships on the basis of their research rather than teaching; the two correlate but with a lot of noise). I’d guess this variance is much greater than institutional variance, though obviously I can’t say for sure.
Also, your phrasing suggested that you thought Harvard had very high-quality teaching. It’s probably good, but I wouldn’t be surprised if most non-research colleges had higher average teaching quality, at least for college-level courses that don’t require deep knowledge of highly advanced material. I’d guess that the advantage of taking classes at Harvard is more in the peer group (see other comment), and maybe the advanced undergrad or grad-level courses where command of the subject matter might vary more. (However, I’ve only ever taken classes at Harvard, so someone who’s transferred in or out would have better info.)
Also, students here weigh instructor quite heavily in their course decisions, at least in the math department where professors switch courses almost every year.
Oh, sorry. I think I interpreted “assuming that the major and course level are given” differently than you intended. My university is too small to offer different sections of the same course by different instructors.
EDIT: I see that CarlShulman brought up essentially the same points, which I somehow managed to miss while posting this. I’ve left up the original comment for posterity but feel free to ignore.
I have some questions about this.
First, at least for the example small-probability failure modes that you gave as an example in your previous article, an individual worker would not be in a much better position to assess them than a donor (or at least, would be in no better position to assess them than a donor’s inputs, like GiveWell or other third parties). Can you give some examples where workers would be in a better position?
Second, donors seem to be in a much better position than workers to react to failure. If you’re doing something that hasn’t been done before—which seems to be the place where direct work is most obviously better than earning to give—then you need to spend a lot of time figuring out things that nobody else has figured out yet. This means that the costs of switching to a different cause are quite high. On the other hand, a donor can simply change their beneficiary organization, which is much easier.