Advice for a Budding Rationalist

by atucker1 min read19th Nov 201027 comments


Personal Blog

Most people in the US with internet connections who are reading this site will at some point in their lives graduate high school. I haven't yet, and it seems like what I do afterwards will have a pretty big effect on the rest of my life.* 

Given that, I think I should ask for some advice.

Any advice? Anything you wish you knew? Disagreement with the premise? (If you disagree, please explain what to do anyway.)

More specific to the site,
Any advice for high schoolers with a rationalist and singularitarian bent? Who are probably looking at going to college?
Anything particularly effective for working against existential risk?
Any fields particularly useful for rationalists to know?
Any fields in which rationalists would be particularly helpful?

This is intended to be a pretty general reference for life advice for the young ones among us. With a college selection bent, probably. If you're in high school and have a specific situation that you want help with/advice for, please reply to this post with that. I think that a most people have specific skills/background they could leverage, so a one-size-fits all approach seems to be somewhat simplistic.

*I understand that I can always change plans later, but there are many many things that seem to require some level of commitment, like college.

As Unnamed pointed out, also look at this article about undergraduate course selection.

27 comments, sorted by Highlighting new comments since Today at 8:23 PM
New Comment
[-][anonymous]11y 14

What I wish I'd known:

College is disorienting. It's very different. Some people (I was one of them) initially get really overwhelmed by exposure to the wider world: things like managing your own time, making judgments independent of your parents', encountering other people's cruelty, dealing with relationships, deciding among different professional/academic interests, staying hopeful and pragmatic despite tough competition. Growing up is tough and confusing and I think usually involves a lot of missteps and angst. I'm only just out of college and I certainly haven't got it all figured out.

But my advice is to take care of yourself -- do things that help you stay resilient despite all the confusion. First, get enough sleep, a nutritious diet, some exercise, and again ENOUGH SLEEP. Second, friendship is incredibly valuable, especially the friendship of stable, successful people you can look up to. The time you spend on friendship isn't wasted. Third, never miss deadlines and never skip class. Old-fashioned rules like that are a bulwark. You may be experimenting with new activities, or having personal drama, or going through a tough time: but if you never miss deadlines, you're quite a bit safer from real catastrophe. Fourth, nobody ever died from knowing too much math. Put some attention into learning things that you can use to support yourself in the future. This doesn't mean you should pick a "safe" major -- there are no safe majors -- but plan for the future, and acquire quantitative skills no matter what. Fifth, love yourself. Hokey, but very, very real.

nobody ever died from knowing too much math.


I hope I'm able to follow through on the taking care of myself bit. Its sorta annoying how many people know stuff like that, but don't actually do it.

Slight question: What kinds of math do you find cool/helpful? I know a lot of calculus, a good deal of statistics, basic probability stuff, and a smattering of information theory (enough to not get totally lost on wikipedia). Am I missing any? Linear Algebra comes to mind.

[-][anonymous]11y 1

It depends what you want to do with yourself: if you have a technical major, the mathematical requirements will become clearer. (For example, graph theory is much more important in computer science than in, say, physics.) For general knowledge, learn linear algebra and multivariable calculus. If you haven't already, read a book/take a class that's heavily proof-based -- for example, Rudin's Principles of Mathematical Analysis. Being comfortable with formal proof is necessary for a number of disciplines (not just math) and is, I think, good for your brain. (I think of learning to prove as a similar kind of transformative experience as learning to read -- except that nearly all of us learn to read and not so many of us learn to prove.)


I cannot believe I forgot to mention this. This is immensely good advice.

You can stay and should up to 3 AM talking to friends about deep and interesting things, but you should never require all-nighters to get your work done.

never skip class

A slightly refined suggestion: have perfect attendance for the first month and a half, and then only skip if you've already worked ahead as much as possible. You don't want to judge the value of a class by its introduction or review period, and the only thing worse than wasting time in a trivial but necessary class is doing poorly in a trivial but necessary class.

[-][anonymous]11y 0

Expanding on the sleep thing:

I thought it was an old wives' tale/not applicable to me. But when I actually tried the experiment, documenting how I felt on 4 hours of sleep vs. 8 hours of sleep, I realized that I was happier and more intelligent when rested, and that I understood lectures much better when I didn't doze off! If you doubt that this is applicable to you, try the experiment.

I never actually needed all-nighters to finish my work. If you space out your studying it's never necessary.

Read a bunch of Paul Graham.

While in high school: treat it like a day job. Work on your interests, and develop your skills.

Try to avoid video games. Avoid alcohol.

In college: Don't focus on classes. Focus on people, and focus on ideas. Your goal in college should be to meet a few people cleverer than you are (settle for as clever if necessary). Introduce yourself to people, and don't get locked into the group that forms in the first week. If you enjoy a professor's class, go to their office hours and chat. Remember, understanding is not something you finish- it is the shore of an island. The more you know, the more you are immediately able to learn- and so if you've got what's in the curriculum find one what's on the edge of the curriculum.

It seems to me very likely that the best way to reduce x-risk is to work on other issues you're better at and then donate the money you're able to earn there. "If you want to create wealth (in the narrow technical sense of not starving) then you should be especially skeptical about any plan that centers on things you like doing. That is where your idea of what's valuable is least likely to coincide with other people's." --Paul Graham

Math seems like a pretty foundational field for a lot of things. I recommend becoming a competent programmer in at least one language (preferably something like Ruby, but you can probably make more money freelancing with something like Perl).

Don't be afraid to seize value. If there are classes you find interesting that you aren't in, go to them. If there are professors you want to talk to about things, go to their office hours, or email them, even if you aren't in their class or their department.

I avoided alcohol for the first 30 years of my life.

I now have about 7 beers (or equivalent) per week (never more than 3/day). There's a temporary impairment for sure, and sleep suffers with several drinks just before, but I would definitely not recommend against moderate drinking.

[-][anonymous]11y 4

I'm not opposed to drinking -- I drank socially in college, and I knew many successful people who drank far more. What's more dangerous is overwhelming peer pressure. I've known people who joined clubs where lots of drinking was all but mandatory, and their lives got a lot more chaotic than they wanted, very fast. Keep away from mandatory partying; discretionary partying, on your own schedule, with people who will respect your choices, is much better.

You're right. I did get that kind of pressure at college parties. Really, people will be happy as long as you have one drink; the further teasing and egging on is just some stupid ritual program that doesn't mean anything to them. It's definitely weird to put yourself in a situation where people are gathered to drink+have fun, and not drink at all.

This thread is advice for people under ~22 and new to both adulthood and choosing friends rather than lucking into them. It seems that avoiding alcohol is a strongly positive idea in such a situation. (Especially if, like the OP, one is worried about x-risk.)

Afterwards, I trust them to make good decisions, and that can very well include alcohol (just like good decisions can very well include playing video games).

Clearly getting dangerously drunk or becoming an alcoholic are things which young people should avoid. As such caution against drinking is good advice for lots of young people. However, young Less Wrongers are more likely to be lonely, isolated and not having as much fun as they should. Going through college without drinking is likely to make dealing with these issues dramatically more difficult as so much socialization revolves around alcohol.

However, young Less Wrongers are more likely to be lonely, isolated and not having as much fun as they should. Going through college without drinking is likely to make dealing with these issues dramatically more difficult as so much socialization revolves around alcohol.

My suspicion is that drinking buddies are generally a net negative, as they may stop you from looking for other friends. My fallacious guess is most LWers are satisfied socially once they have a few friends they meet with regularly- and so if you find drinking buddies to tide you over until you find solid friends, you may find yourself looking for solid friends less and less.

I suspect that advice is better given as generally applicable, than as generally applicable but there are caveats. For one thing, policy debates should be two-sided, and for another people are simply bad at telling when the rules don't apply. Anyone reading this is hopefully clever enough to take my injunction against alcohol and weigh it against the other information going into their decision; if they aren't, I think it's a good plan for them to avoid alcohol entirely. You are right that adding the motivation behind my injunction helps make it more potent, and possibly limits it in a good way.

[edit] One more thing- if you need alcohol to warm up and talk to people, what you really need is to get better at talking to people. That is a skill you can practice and depends on confidence you can develop.

I would only advise that a person choose rationally how much alcohol to consume (i.e. gain experience predicting the effects and aftermath of drinking, and make decisions about how much to drink before becoming inebriated). I don't see any strong evidence that moderate and steady (as opposed to binge) drinking is helpful or harmful to health.

If someone wants to always optimize their rationality-in-the-moment and has no social or emotional needs, then I guess it's safe to never get drunk. It certainly hasn't helped me become smarter or more rational to do so. I just think it's stupid from a hedonistic perspective to absolutely avoid it.

Thanks, I particularly found the whole "just go to classes you feel like" bit interesting. Slash looking forward to it.

So for most people directly working on existential risk is inefficient. How can you tell if it would be efficient?

So for most people directly working on existential risk is inefficient. How can you tell if it would be efficient?

This question is the heart of why I suggest avoiding it, as a field. Feedback is golden. Imagine writing something where, as soon as you pressed a button, it would give you feedback on what you've written, compared to writing something where you had to print it out, mail it out, wait for a reply, and then get feedback. Even if the feedback is ten times better in the second case, the first one will encourage you to work much harder and faster.

So, when it comes to x-risk it seems the best plans are ones that increase feedback. If you want to improve our systems to guard against asteroids striking the Earth and catastrophic climate change, it seems like the best approach is to acquire millions of dollars and create a satellite network that generates better information about what Earth's weather is like and tracks near-Earth asteroids. (And, the first approach in such a plan is to call up your Congressperson and a NASA head, trying to arrange a private grant, rather than doing this yourself with your engineering buddies. If you do decide it's better to do it yourself, poach experienced people, don't start with inexperienced ones.) Information is generally a net positive, while opinions are not nearly as valuable.

So, if your choice is "do I go into a value-generating field, or do I write papers about how scary asteroids are?" it seems to me the first is the better choice. If no one is approaching the issue you are interested in, then maybe you should devote yourself to it.

Would an appropriate short version be "Doing stuff takes money and data, so you should be working towards one of those"?

I think that's a good summary of half of it. The other half is opinions are high cost and low value for most applications.

Be attractive and popular. We need rationalists who will understand the need for x-risk avoidance and be able to get the humans to do whatever they need to do to save themselves. It will also help you get money, which can be used to buy happiness.

With some rare exceptions that test both, most college courses test either your intelligence or your diligence. A diligence course requires you to memorize material you will never care about and read books you're not interested in. An intelligence course will give you a few models to manipulate or a new type of math to master. If you're good with numbers and what-if scenarios (i.e. supply/demand clicked quickly for you, you enjoyed calculus), then I'd definitely recommend taking intelligence type courses. While your peers spend hours studying, you can read about other subjects you're interested in. If you have the hardware, it's a relatively painless way to slip through one of life's most expensive hoops.

I'm about to graduate with a major in Econ and a minor in math. I'd do it again in a heartbeat. My econ electives were mostly spent on statistics and game theory. Look for classes like those. Stay away from anything with history in the name.

The intelligence/diligence split should be obvious by looking at the class, right?

Why do you say to avoid history?

History classes, especially in college, tend to be large lectures with the instructor running a powerpoint, while students take down notes. Exams are infrequent and will occasionally include an essay that may require a bit of critical thinking, but the overwhelming majority of the work just goes into absorbing information. I happen to like history, and it's not too difficult because I have a good memory, but it certainly isn't really a thinking class. Actually, the best history class I've had was fairly small, but we spent most of the class time discussing what was going on, instead of having it just being dumped on us. Most history classes aren't like that though.

This is pretty specific, but a thing I wish I had done when studying is to try spaced repetition software seriously for learning the course material.

Assuming from the start that teaching is probably useless, I'm responsible for figuring out good ways to learn the material, and the faculty is there mostly to specify the course contents and test that I've ended up learning them correctly might have helped. I found undergraduate math lectures completely unfollowable, but was able to figure out most things from working through textbooks with pen and paper. I still suspect that lecture-centered courses on any topic involving learning new formalisms are some kind of historical mistake, but ymmv.

  1. Don't be afraid to take the highest level classes in high-school, if you get the chance. In my experience people in higher level honors classes tend to be nicer and more tolerant of individuality. The subject matter may be harder, but homework is often lighter, (and the other students less disruptive) and that gives you more time to study. If your school gives extra weight to honors/AP classes when determining GPA, the decision should be a no-brainer.

  2. Similarly, people in college are genuinely nicer than high school students. Partially, this is just because college students are more mature, and have learned self-control more than high-schoolers. But a big part is the selection effect, the more conscientious/intelligent students are nicer, and they go on to college more often than others. Plus, of course, most college students want to be there. Lots of teenagers see their high schools as being like prisons, and they are resentful. I witnessed kids getting bullied every day in high school. I never saw that in college (and I've spent time in 3 schools, counting grad school). So you can be yourself more at college than in high school: take advantage of this!

  3. I wish I had spent more time as an undergraduate going to parties, drinking, and doing internships/ independent research. I wish I had spent less time studying, surfing the internet, and playing video games with roommates.

  4. I drove around the US with a buddy one summer, visiting national parks and beaches. I wish I had done more of that and not rushed to graduate early.

  5. Majoring in something like history or philosophy is fine but don't bother going to grad school in that subject area (return on investment for M.A. degrees and PhDs are much lower than for BA degrees). Majoring in something where you actually learn useful knowledge or skills (like computer programming or engineering) is only useful if you actually want to work in those fields.

Playing video games with roommates was the best time I had in college. Until said roommates decided that someone else would take my place in the room next year. :(

I actually ended up having a much better time overall in high school than in college.

Some parts of this thread about advice for an undergraduate should be relevant for you too.

Thanks a lot! I think this is the one I saw a while ago.