College advice

This post is a collection of several blogposts about college advice, or advice not directly about college but germane to it. Each post is worth reading in full, though I have added excerpts.

To start with, Ben Kuhn’s college advice for people exactly like him:

Get better at deciding what to do

One thing I wish I’d had more of in college is something like a sense of taste: not aesthetic taste, but more “taste for what’s awesome” or “taste in good projects.” (My role model for good project-taste was my friend Adrian, who coauthored articles in Nature, Science and Cell before graduating.)

Good taste is what will guide you to doing effective things, instead of pointless things, in college. But it’s also really hard to develop good taste while in college, because you have no models of good taste except for academics. Academics’ taste is often deeply weird and driven by what niche will get them tenure rather than what is exciting or useful, so it’s good to have other models of taste too.

My best guess for how to do that is to take a gap year, during which you can front-load learning how the real world works. Summer internships outside academia are also good. [1]

Also from Kuhn’s piece: “if your college assigns you advisers they will probably be terrible”. I don’t know if I’d put it quite that way, but there’s definitely value in thinking for yourself and trying to get around well-intentioned but pointless bureaucracy.

I have a much easier time endorsing this from his post, though:

Aggressively ignore bullshit

Coming in, you will probably think that your college experience has been carefully designed to provide you the best possible education. (HA HA HA. [3]) As a result, when your college asks you to do things, you will feel inclined to take them seriously.

This is a mistake. Your college will spew lots of bullshit at you. Ignore it when you can; when you can’t, try to limit the degree to which it seeps into your life.

Examples of things that, in my opinion, turned out to be bullshit (of course this will vary by school, field, worldview, etc.):

Physically attending lectures (just listen to them at 2x speed)

“general education”

Mandatory writing seminar (my writing got worse, not better)

Paperwork (let me tell you the story of why I don’t have a minor in computer science!)

Prerequisites for courses (note: not always!)

Many courses (but very much not always!)

Lest this sound too cynical, let me point out that (a) college is still awesome and (b) ignoring bullshit gives you crazy superpowers to focus on the actually awesome parts, like spending time with smart people, learning difficult things, trying out different jobs and activities, etc.

Consider speedrunning college. From Claire Zabel’s talk:

There are great reasons to get a university degree. It opens up a lot of different options to you. But maybe you should try to get it as fast as possible. I think if you try to take easy classes, you might be able to potentially get a degree a year faster.

… [S]ome people want to pursue strategies where the stuff you learn at university just doesn’t help that much. It’s good to recognize that and be aware of it, especially if you have good fallback plans and easy-to-demonstrate skills, or if you just hate being at university. I know a lot of people who really don’t enjoy their time at university. It feels competitive. It’s very fast-paced; things change all of the time. You have to sit and listen to people a lot when you might prefer doing things. It can be pretty valuable to move on to a different phase of life.

Related to this: Julia Steinberg’s “Don’t Coterm”: “Coterming would mean stapling a year-long master’s degree program to your regular B.A. or B.S., usually taking the form of a fifth year. … [T]he vast majority of students who coterm use these thinly veiled excuses to disguise their allergy to adulthood.” And Isaak Freeman’s “University & Adversity”: “I did a 4-year program at in 2 years. … Figuring out how to get around any obstacles, usually by asking, then getting rejected, then staying persistent until it worked out”. See Tego’s post for a granular how-to on graduating early.

Also relevant here is Derek Sivers’ “There’s no speed limit”, an account of how he covered six semesters of a music undergraduate degree in five three-hour lessons:

Kimo’s high expectations set a new pace for me. He taught me that “the standard pace is for chumps” — that the system is designed so anyone can keep up. If you’re more driven than most people, you can do way more than anyone expects. And this principle applies to all of life, not just school.

Read Xiaoyu He’s “Yoda Timers 3: Speed”:

Everyone has a rough idea of how long things have to take. Solving a hard research problem always takes at least a month, right? Writing a paper should take at least an hour, right?

When I first started playing the Arithmetic Game for middle school MathCounts training, my high score was close to 20. After a few months of dedicated training, my record hit 90, making it onto the leader-board of the time.

For any given task, do not assume you’re doing it anywhere close to your real speed limit. It used to take me at least four hours to write a blog post this length. This one clocks out in just under forty minutes.

On that note, also read Patrick Collison’s “Fast”, a list of “examples of people quickly accomplishing ambitious things together”.

James Somers’ “Speed matters: Why working quickly is more important than it seems”:

The obvious benefit to working quickly is that you’ll finish more stuff per unit time. But there’s more to it than that. If you work quickly, the cost of doing something new will seem lower in your mind. So you’ll be inclined to do more.

From Ben Kuhn’s “Be impatient”:

There’s an obvious way in which moving faster is important: if you’re 10% more productive, you will finish your work in 10% less time, so you can do 10% more work total. But I don’t think that’s the main reason that speed is important.

It’s worth pointing out at this point that all of the quotes above aren’t just about churning out work—they’re about processing information more quickly. The faster you process information, the faster you can incorporate the result into what you do next.

In other words, the main benefit of being fast is that you end up doing different things.

From Nat Friedman’s list of things he believes:

It’s important to do things fast

You learn more per unit time because you make contact with reality more frequently

Going fast makes you focus on what’s important; there’s no time for bullshit

“Slow is fake”

A week is 2% of the year

Time is the denominator

Read Stephen Malina’s “Energetic Aliens” for vivid examples of work intensity.

I hold the probably unpopular opinion that you can learn to like almost any field of study, once you have devoted enough hours to mastering it. This is because i) there are interesting questions at the frontiers of almost all subjects, and ii) mastery intrinsically feels great. But expect the initial hours to be painful slogging. As Sivers puts it in a different blogpost: “The hours don’t suddenly appear. You have to steal them from comfort. Whatever you were doing before was comfortable. This is not. This will be really uncomfortable.” You will need to do time in what Cate Hall calls the moat of low status. Flow state is incredibly satisfying, but it’s also rare. Lots of deliberate practice doesn’t feel like flow state, yet it’s valuable.

Where does Sivers’ speedrunning fit into all this? I think that if you can, you should chunk these painful hours together. Go through them over a few days instead of a few years. That way, you notice that you’re making fast progress, which is motivating. Speedrunning a subject also pushes it near the top of your mind, which means your idle hours get devoted to it.

The talk Sivers gave to incoming first-year students at the Berklee College of Music is also interesting:

While you’re here, presidents will change, the world will change, and the media will try to convince you how important it all is.

But it’s not. None of it matters to you now.

You are being tested.

Your enemy is distraction.

Stay offline. Shut off your computer. Stay in the shed.

When you emerge in a few years, you can ask someone what you missed, and you’ll find it can be summed up in a few minutes.

The rest was noise you’ll be proud you avoided.

See Warren Buffett’s “2 List” scheme for deciding what to focus on:

STEP 1: Buffett started by asking Flint to write down his top 25 career goals. So, Flint took some time and wrote them down. (Note: you could also complete this exercise with goals for a shorter timeline. For example, write down the top 25 things you want to accomplish this week.)

STEP 2: Then, Buffett asked Flint to review his list and circle his top 5 goals. Again, Flint took some time, made his way through the list, and eventually decided on his 5 most important goals.

Note: If you’re following along at home, pause right now and do these first two steps before moving on to Step 3.

STEP 3: At this point, Flint had two lists. The 5 items he had circled were List A and the 20 items he had not circled were List B.

Flint confirmed that he would start working on his top 5 goals right away. And that’s when Buffett asked him about the second list, “And what about the ones you didn’t circle?”

Flint replied, “Well, the top 5 are my primary focus, but the other 20 come in a close second. They are still important so I’ll work on those intermittently as I see fit. They are not as urgent, but I still plan to give them a dedicated effort.”

To which Buffett replied, “No. You’ve got it wrong, Mike. Everything you didn’t circle just became your Avoid-At-All-Cost list. No matter what, these things get no attention from you until you’ve succeeded with your top 5.”

This tweet thread is also great:

A thing econ/finance PhD students sometimes do is overload up on stuff like TA/RA-ships, reading groups, attending seminars, etc. etc… until the amount of work time they have left over to actually write papers is very low. Don’t forget, your goal during PhD is to write papers!

On the job market, nobody directly cares about how well you did on your reading group presentation or mock referee report or problem set, close to the only thing that matters is whether you wrote good papers

There are usually only a handful of things that really matter!

On setting aside time for focused work, see this from Leila Clark’s “101 things I would tell my self from 10 years ago”:

Start tracking hours spent on deep work. It is an obvious and very visible metric of how much work you have done.

I also recommend Cal Newport’s book Deep Work. Most self-help books could be blogposts. Deep Work was valuable cover to cover.

shaurya’s tweet:

you have access to your optimal policy at all times. you just choose not to follow it. you can literally access it by asking “what should i be doing rn?”

David Holz’s tweet:

How to find a job in Silicon Valley:

1) Ask yourself which companies have missions that you’re most passionate about 2) Think about how you can help 3) Do a few projects that show you can help (ie: coding, design, marketing concepts, educational blogposts, documentation, as long as it’s relevant to the work / mission / company) 4) Reach out to the founders or senior employees directly (not via recruiting channels) and say “I’m passionate about your mission X for Y reasons. I think I can help by doing Z. Here are some projects that show I can do it.”

If it doesn’t work out, don’t be discouraged. Ask for feedback. Reach out to more companies and founders that you’re passionate about helping. Do more projects that show off your strengths. You’ll make it.

You don’t need an education or resume for any of this. Your projects, determination, and research are your education + resume.

Allen Naliath’s tweet:

Here’s my advice on how to guarantee an internship/job

Now that we’ve announced @fridaymail, I’ve gotten over 100 requests asking for a job/internship

95% are instantly rejected

Why?

Because they’re all missing the one thing that’ll convince me (or any employer) to hire you:

Send the employer a live, working link to the most impressive project you can build

Milan Cvitkovic’s “Things you’re allowed to do”:

Ignore what’s on the jobs page and directly pitch someone at a company on hiring you

The jobs page is always out-of-date anyway

Figure out what their needs are before you make your pitch

Byrne Hobart’s “In Defense of Hiring Based on Side Projects”:

Making a habit out of constantly improving the craft is good because it speeds up your execution of known tasks (“Figure out why this button looks weird in IE,” “Update this financial model to account for the customer they just lost,”) but also because it helps reduce the Knightian uncertainty around what problem to solve.

Nick Cammarata’s tweet:

i hate how well asking myself “if i had 10x the agency i have what would i do” works

Andrej Karpathy’s “Doing well in your courses”:

All-nighters are not worth it.

Sleep does wonders. Optimal sleep time for me is around 7.5 hours, with an absolute minimum of around 4hrs.

It has happened to me several times that I was stuck on some problem for an hour in the night, but was able to solve it in 5 minutes in the morning. I feel like the brain “commits” a lot of shaky short-term memories to stable long-term memories during the night. I try to start studying for any big tests well in advance (several days), even if for short periods of time, to maximize the number of nights that my brain gets for the material.

I think that “maximizing the number of nights that your brain gets for the material” is a wonderful phrase.

NEVER. EVER. EVER. Leave test early.

You made a silly mistake (I guarantee it), find it and fix it. If you can’t find it, try harder until time runs out. If you are VERY certain of no mistakes, work on making test more legible and easier to mark. Erase garbage, box in answers, add steps to proofs, etc.

I have no other way of putting this — people who leave tests early are stupid. This is a clear example of a situation where potential benefits completely outweigh the cost.

This is good advice, but I have left tests early before, and I don’t think I am stupid. (I was taking those classes pass/fail, or the test impacted my grade by too little. So, arguably those exceptions prove the rule.)

Patrick Collison’s advice for people aged 10-20:

To the extent that you enjoy working hard, do. Subject to that constraint, it’s not clear that the returns to effort ever diminish substantially. If you’re lucky enough to enjoy it a lot, be grateful and take full advantage!

On the importance of working hard, here’s Byrne Hobart:

I think there’s a tendency for people to underestimate just how much continuous effort is behind really big successes.

There are natural talents, of course, but what natural talent does is

1) have an effect on what your limit is if you invest maximum effort, and 2) give you rewarding feedback right when you start something, which is a good reason to keep at it.

At the tails of the distribution, talent is the differentiator, but most of us are not in those tails, so most of us should work harder. And at the tails, it still matters; there are plenty of talented people who never did anything besides notice that they were talented.

I think it’s especially important to work hard if you’re near the top of the distribution in natural talent. This is because you can be convinced that your output per hour is the highest possible. The rate at which you’re learning/doing something is the fastest someone can learn/do that thing. So, if you work longer than everyone else, you’re sure to produce more than anyone else. This matters because there are often superlinear rewards for output. The best search engine doesn’t just get the most users; it gets effectively all the users. The first historian or economist to mine a new data source doesn’t just get a proportionally larger share of the citations that could possibly result from that source; they get all the citations. If you’re convinced you’re an outlier natural talent, cranking the effort dial to 11 is a surefire way to get outsized success. And as dials go, effort is a straightforward (if not easy) one to turn.

Also on the importance of working hard, see this from Paul Graham’s “How to Work Hard” (which is a great essay and worth reading on its own):

One thing I know is that if you want to do great things, you’ll have to work very hard. I wasn’t sure of that as a kid. Schoolwork varied in difficulty; one didn’t always have to work super hard to do well. And some of the things famous adults did, they seemed to do almost effortlessly. Was there, perhaps, some way to evade hard work through sheer brilliance? Now I know the answer to that question. There isn’t.

The reason some subjects seemed easy was that my school had low standards. And the reason famous adults seemed to do things effortlessly was years of practice; they made it look easy.

Gwern’s “On Really Trying” is a truly stunning collection of people showcasing extraordinary motivation—often, simply after hearing what they once thought was impossible had been done. I’ll highlight two anecdotes.

From Walter Isaacson’s Steve Jobs:

One of Bill Atkinson’s amazing feats (which we are so accustomed to nowadays that we rarely marvel at it) was to allow the windows on a screen to overlap so that the “top” one clipped into the ones “below” it. Atkinson made it possible to move these windows around, just like shuffling papers on a desk, with those below becoming visible or hidden as you moved the top ones. Of course, on a computer screen there are no layers of pixels underneath the pixels that you see, so there are no windows actually lurking underneath the ones that appear to be on top. To create the illusion of overlapping windows requires complex coding that involves what are called “regions.” Atkinson pushed himself to make this trick work because he thought he had seen this capability during his visit to Xerox PARC. In fact the folks at PARC had never accomplished it, and they later told him they were amazed that he had done so. “I got a feeling for the empowering aspect of naïveté”, Atkinson said. “Because I didn’t know it couldn’t be done, I was enabled to do it.” He was working so hard that one morning, in a daze, he drove his Corvette into a parked truck and nearly killed himself. Jobs immediately drove to the hospital to see him. “We were pretty worried about you”, he said when Atkinson regained consciousness. Atkinson gave him a pained smile and replied, “Don’t worry, I still remember regions.”

From an NYT article titled “That Which Does Not Kill Me Makes Me Stranger”:

In all decisions, Stanovnik governs according to a rule of thumb that he has developed over the years: at the dark moment when Robič feels utterly exhausted, when he is so empty and sleep-deprived that he feels as if he might literally die on the bike, he actually has 50% more energy to give. …

Interestingly—or unnervingly, depending on how you look at it—some researchers are uncovering evidence that Stanovnik’s rule of thumb might be right. A spate of recent studies has contributed to growing support for the notion that the origins and controls of fatigue lie partly, if not mostly, within the brain and the central nervous system. The new research puts fresh weight to the hoary coaching cliché: you only think you’re tired.

Nietzsche’s “will to power” and Larry Page’s “healthy disregard for the impossible” are great phrases.

Henry Oliver’s “How do the most prolific people produce so much?”:

Ego. This sounds obvious, but someone like Frank Lloyd Wright was driven by a monstrous sense that he had a mission to save architecture, which is no small part of why more than half his life’s work was done after retirement age. Virginia Woolf, prolific considering her early death (include her diaries) was similar. Noel Coward, who could have been more productive but still produced an awful lot, shares with them a childhood that, directly or indirectly, nurtured a sense of excellence: they had to be the ones to do this. It was part of living up to their idea of themselves.

Sandy Maguire’s “You Don’t Need to Be Brilliant to Do Brilliant Work”:

There likely are problems out there that are brilliance-constrained, but I’d argue that there are 100x more problems which are merely effort-constrained. This is good news, because while it’s not clear how to become smarter, it’s very doable to just throw more energy at something.

It’s the bitter lesson of productivity: you’ve just got to throw more compute at the thing.

See also Maguire’s “No Miracle People”:

The key takeaway from all of this is that your mind is better considered as a blacksmith’s shop than as a pristine sculpture. The blacksmith constructs his own tools, with which he can construct better refined goods and even more precise tools. Your mind isn’t a precious artifact passed down from generation to generation whose flaws and imperfections you have no choice but to deal with. It’s a living, breathing thing, with the capabilities to shape itself into whatever you want it to be.

Related to this, the efficient markets hypothesis holds in only very specific circumstances. As cheesy as quoting a billionaire is, the world seems courage-constrained; there are $10 bills on the ground everywhere for those with the eyes to see. For example, one of my friends new to open-source development corrected a majorly and pointlessly bad KL divergence estimate in HuggingFace’s TRL library that had gone unnoticed despite 13.1k stars and 1.8k forks. A random smart person fixed COVID vaccine tracking in California, a hobbyist discovered the only known aperiodic monotile, etc.

One of my mentors put it this way: “Hard problems and easy problems actually take the same amount of time.” See Ben Kuhn’s “You don’t need to work on hard problems”:

I won’t claim that the returns to intelligence are literally equal everywhere. But given a hairy enough goal, they’re a lot more evenly distributed than you think. So don’t look for hard problems—important ones are ultimately more fun!

The breadth of John Wentworth’s writing impresses me. He has a study guide for (technical) topics to study to become a generalist problem-solver. It also contains some good college advice. For example, I wholeheartedly agree with this:

[D]on’t just take whatever courses are readily available. I recommend heavy use of online course material from other schools, as well as textbooks. Sometimes the best sources are a lot better than the typical source

and this:

I’ve found it useful to “pregame” the material even for my normal college courses - i.e. find a book or set of lectures covering similar material, and go through them before the semester starts, so that the in-person class is a second exposure rather than a first exposure.

Wentworth recommends breadth over depth so that you are, at least in principle, able to model anything. But picking something to specialize in is also useful, as this comment points out:

I’d adjust the “breadth over depth” maxim in one particular way: Pick one (maybe two or three, but few) small-ish sub-fields / topics to go through in depth, taking them to an extreme. Past a certain point, something funny tends to happen, where what’s normally perceived as boundaries starts to warp and the whole space suddenly looks completely different.

When doing this, the goal is to observe that “funny shift” and the “shape” of that change as good as you can, to identify the signs of it and get as good a feeling for it as you can. I believe that being able to (at least sometimes) notice when that’s about to happen has been quite valuable for me, and I suspect it would be useful for AI and general rat topics too.

It is for similarly meta reasons that Holden Karnofsky recommends doing minimal-trust investigations: deep dives in which you “suspend my trust in others and dig as deeply into a question as I can”. You become far more knowledgeable about whatever you’re investigating after a minimal-trust investigation, but the primary value is helping “develop intuitions for what/whom/when/why to trust” and getting “a demonstration and reminder of just how much work minimal-trust investigations take, and just how much I have to rely on trust to get by in the world”.

LessWrong’s Starting University Advice Repository has some good comments, like MaximumLiberty’s — “Audit classes that you wish you had time to take.” Professors (generally) love students who’re attending their classes just for fun.

Vaniver on approaching your social life with agency:

In many residential universities, there’s a tremendous amount of social inertia—the people you’re friends with a year or two in will often depend heavily on who you met your first few weeks, and who you met in the first weeks of classes. So turn this to your advantage: introduce yourself to people you don’t know early on, and try to deliberately figure out who’s a good fit and who’s a bad fit instead of just trusting to chance.

In a different blogpost, William Eden echoes this “social circles form quickly” point:

In terms of who you spend your time with, recall what I said about the default circumstance… you hang out with the freshmen on your floor. These people only have one comparative advantage: they are right next to you. Admittedly it can be really nice to have immediate social opportunities with low transaction costs, so be conscious about that factor. Now, the people you will truly click with will likely be distributed all across campus. In this case, time is of the essence — people’s social circles become ossified quickly, so it becomes more and more difficult to establish a strong connection over time. In this case initial exploration is the right strategy. Think about your interests, find out where those folks would hang out, look into lots of groups, go to parties and meet people. Ask other people who they think you should meet.

A significant value prop of college is these connections. See Siderea’s “The Value of College”:

There was a famous paper written by a guy named Granovetter about how people find jobs. It was called “The Strength of Weak Ties”. He found out that it’s not your BFFs who help you find work, because you already know all the same job leads they know. It’s the people you are friendly acquaintances with who get you job leads. They are well disposed to you (being friendly) and they know things you don’t, being as they are enmeshed in different families and friendship groups than you are.

At the same time, it’s counterproductive to think of college as “networking”.

Also see Manfred’s comment from the repository:

Research as an undergrad can be very important, but on average it’s a waste of time - don’t do it just to do it. In order to make it work you have to be above average in finding out what research is currently going on in the field you’re interested in and at your university, and pick out groups doing research you are interested in and using methods that you are interested in learning about.

Max Slater’s review of CMU courses contains some advice for freshmen. See particularly this bit about undergrad research:

Immediately doing research isn’t the best idea: even if a professor gives you a real project, your knowledge will be pretty limited until you’re able to take CS electives. I would instead recommend focusing on doing excellently in/more of the core courses. Over the following summer, however, getting into research is a great idea if you don’t want/get an internship, since first year internships are unusual.

Jason Eisner has a collection of advice for research students. Here’s his advice for undergrad research:

Either you make it your top priority, or you don’t do it at all. … Your research advisor doesn’t get much credit for working with junior students, and would find it easier and safer to work with senior students. It’s just that someone gave her a chance once: that’s how she ended up where she is today. She’d like to pay that debt forward.

A hack to make research your top priority is to actually enjoy it. And the way to enjoy it is to be personally curious about the research question: “[imagine] that another group has published the paper you have in mind. Are you excited to read it?”

Henrik Karlsson’s “A funny thing about curiosity”:

[T]he path of maximal interestingness is supposed to feel like fun. Not fun as in “I feel entertained” but fun as in, “this is engrossing and self-surprising, life-affirming and a little scary.”

Or be “guided by beauty”, as Jim Simons puts it:

Be guided by beauty. I really mean that. Pretty much everything I’ve done has had an aesthetic component, at least to me. Now you might think ‘well, building a company that’s trading bonds, what’s so aesthetic about that?’ But, what’s aesthetic about it is doing it right. Getting the right kind of people, and approaching the problem, and doing it right […] it’s a beautiful thing to do something right.

Simplicity is an underrated source of beauty. “Simplicity is the ultimate sophistication,” da Vinci (never) said. Many great ideas are obvious in hindsight. “If aliens did natural language processing, they’d probably use attention,” one of my ML professors said. “How extremely stupid not to have thought of that!” said Thomas Huxley when he read Darwin’s On the Origin of Species.

John Schulman’s “An Opinionated Guide to ML Research” is fantastic. Also worth reading: Michael Nielsen’s “Principles of Effective Research”

I’ll highlight this from Tom Silver’s “Lessons from My First 8 Years of Research”:

Find someone who you feel comfortable asking “dumb” questions

I was initially very intimidated by my colleagues and hesitant to ask basic questions that might betray my lack of expertise. It was many months before I felt comfortable enough with a few colleagues to ask questions, and still my questions were carefully formulated. Now I have three or four go-to people. I wish I had found them sooner! Before I was drowning in a backlog of terms to Google after work. Now I immediately ask a question when it comes and my confusion is resolved before it compounds.

It’s understandable, but still extremely insensible, to prioritize looking smart over unblocking yourself to progress as fast as possible. Many months to feel comfortable asking questions is many months too long. Ask questions.

Also note this advice from Silver:

Write!

I once had an occasion to ask a very prominent AI researcher for early career tips. His advice was simple: write! Write blog posts and papers of course, but even more importantly, write down your thoughts throughout the day. Since he said that, I have noticed an obvious difference in progress that I make when I am actively writing versus simply thinking.

Jason Wei’s perspective of research as dopamine management is great. One of my mentors said that biology “humbles you”. I think all disciplines do.

Cold email works well for getting research positions. On cold emailing, see Eugene Vinitsky’s “A Guide to Cold Emailing”:

The problem is that people stress out about cold emailing too much. Yes, people will ignore many or most of your emails (a 10% response rate wouldn’t be unusual). But importantly, no one will hold those emails against you unless you either sound like a huge jerk in your email or get mad at them when they don’t respond. Let me stress this point: either you’ll get a response or they’ll forget you ever emailed them at all. You matter much less than you think. As a consequence of people forgetting, it’s basically something with only upside.

See also Alexey Guzey’s “It Is Your Responsibility to Follow Up”.

Eisner’s “Write the Paper First” is also excellent advice:

Hint #1: You can hack on little sleep, if you know what needs to be hacked. But you can’t write effectively. Writing involves many big and small decisions, which will seem insurmountable when you’re exhausted and panicked.

Hint #2: Five students can be hacking in parallel at the last minute, but they can’t all be co-writing with me at the last minute. You have your own computer but you have to share your advisor.

Jennifer Widom has tips for writing technical papers.

Alex Wang has post on applying to grad school. A point from his post I’ll single out: get fellowships.

Fellowship funds are tied to you, such that professors no longer need to fund you. As such, securing an outside fellowship makes you a much more attractive applicant, as professors would not need to allocate funds to support you (financially). Furthermore, being self-funded buys you a degree of academic freedom: because you are not tied to possibly project-specific or professor-specific funding, you have more latitude to turn down projects that you aren’t interested in and instead work on projects you are interested in.

Byrne Hobart’s “Read.” is great:

Text benefits from two economic coincidences: it has a low marginal cost today, but it used to have a very high one. So older writing is more likely to contain information that smart people with cash on hand thought was absolutely essential, while newer writing is cheaper to produce and easier to search than other kinds of content, so if there’s one specific piece of information you need, you’re more likely to find it in text than anywhere else.

Video is really only good for i) when you’re too tired to read, and for ii) sharing tacit knowledge. (On that note, LessWrong has a list of the best tacit knowledge videos on any subject.)

When you’re deciding what to read, remember: read easier textbooks instead of struggling valiantly. “[E]ven though slogging through tough textbooks makes you feel sophisticated and smart, don’t.” Pain is not the unit of effort. Be “confused about one thing at a time”.

This Richard Hanania post on which three types of books are worth reading is interesting.

Trust that you can learn a surprising lot: polymathy is a myth is a myth. Knowledge compounds, and expertise has a startling convexity.

See Julian Schrittwieser’s productivity advice:

My first recommendation is to learn as much as possible, about the entire stack you work with - in AI, that’s from transistor design and CPU architecture to data centers, internet infrastructure, from assembly to frontend TS. Be able to do any job in the company!

Having the ability to implement and test an idea without needing to convince anyone else that the idea is worthwhile is a productivity multiplier. A solid understanding of all parts of the system also makes it much easier to design algorithms that map well onto the hardware, and to debug tricky problems.

As Jake Seliger says in “How to get coaching, mentoring, and attention”, text is information-dense and reading it gets the points across faster, but actually doing the reading also demonstrates seriousness (that is, if you’re interested in demonstrating seriousness to your professors at all — this isn’t something to do because you “should”):

Once I had a student who said in class that he didn’t like to read fiction. Fair enough; not everyone does and it doesn’t offend me when others don’t share my vices. A week or two later, however, he wanted me to edit his 43 pages of Starcraft fan fiction; when I said that it isn’t possible to be a good writer without being a good reader, he didn’t believe me. Nonetheless I told him that if he read How Fiction Works and discussed it with me, I would read his Starcraft fan fiction. And I would have. He didn’t, of course, and acted like I I had kicked his puppy when I suggested that he prove himself.

Seliger has also written “How Universities Work, or: What I Wish I’d Known Freshman Year: A Guide to American University Life for the Uninitiated”:

I have a theory that virtually everything you learn in universities (and maybe life) is the substance or application of two (or three, depending on how you wish to count) abilities: math and reading/writing. Regardless of what you major in, work on building those two skills.

Seliger quotes Joel Spolsky, who points out that STEM folks should also learn the craft of writing well:

Even on the small scale, when you look at any programming organization, the programmers with the most power and influence are the ones who can write and speak in English clearly, convincingly, and comfortably. Also it helps to be tall, but you can’t do anything about that.

The difference between a tolerable programmer and a great programmer is not how many programming languages they know, and it’s not whether they prefer Python or Java. It’s whether they can communicate their ideas. By persuading other people, they get leverage.

Actually, while we’re at it, that Spolsky post Seliger quotes — “Advice for Computer Science College Students” — is worth reading on its own:

Never underestimate how big a deal your GPA is. Lots and lots of recruiters and hiring managers, myself included, go straight to the GPA when they scan a resume, and we’re not going to apologize for it. …

Why should I, as an employer looking for software developers, care about what grade you got in European History? After all, history is boring. Oh, so, you’re saying I should hire you because you don’t work very hard when the work is boring? Well, there’s boring stuff in programming, too. Every job has its boring moments. And I don’t want to hire people that only want to do the fun stuff.

But I don’t recommend going for the highest GPA you can physically get. Instead, fix a GPA “budget” early in your undergrad — for example, to get into grad school, 3.8 in famously hard schools like UChicago/Princeton and 3.9 in easier schools like Northwestern seems reasonable. A similarly high GPA is probably necessary for law school. (It’s a budget because you can “save” some semesters by studying harder, and “splurge” surpluses away looking for internships or enjoying the spring or, yes, looking for love.) Anything consistently over this pre-determined budget is paying too much attention in class, unless college is unusually good at teaching you the right things. Anything consistently below this GPA budget is bad, and requires either allocating more time to schoolwork or rethinking your GPA target (for example, maybe you got a prestigious internship which reduces the need to grind your GPA for grad school).

Saul Munn’s “Things You’re Allowed to Do: University Edition”:

evaluate not just “will this be good for my career” but “is this among the best options given the limited resources (time, money, energy, etc) that i have”

The post also links to this spreadsheet of grants and fellowship programs, which could be worth checking out.

Sam Altman’s advice for ambitious 19 year olds:

If you stay in college, make sure you learn something worthwhile and work on interesting projects—college is probably the best place to meet people to work with. If you’re really worried you’ll miss some critical social experience by dropping out of college, you should probably stay.

Altman’s productivity advice:

In general, I think it’s good to overcommit a little bit. I find that I generally get done what I take on, and if I have a little bit too much to do it makes me more efficient at everything, which is a way to train to avoid distractions (a great habit to build!). However, overcommitting a lot is disastrous.

Altman’s “How To Be Successful”:

One more thought about working hard: do it at the beginning of your career. Hard work compounds like interest, and the earlier you do it, the more time you have for the benefits to pay off. It’s also easier to work hard when you have fewer other responsibilities, which is frequently but not always the case when you’re young.

Altman’s “The days are long but the decades are short”:

Do new things often. This seems to be really important. Not only does doing new things seem to slow down the perception of time, increase happiness, and keep life interesting, but it seems to prevent people from calcifying in the ways that they think. Aim to do something big, new, and risky every year in your personal and professional life.

Altman’s “What I Wish Someone Had Told Me”:

It is easier for a team to do a hard thing that really matters than to do an easy thing that doesn’t really matter; audacious ideas motivate people.

Altman’s “Researchers and Founders”:

The best people in both groups spend a lot of time reflecting on some version of the Hamming question—”what are the most important problems in your field, and why aren’t you working on them?”

This last quote is a natural segue into recommending Richard Hamming’s “You and Your Research”.

Neil Kakkar’s “Making the most out of college”:

Explore everything that has no direct harmful impact

The sheer effortlessness with which you can explore numerous activities in college is unparalleled. It’s a unique place, where in some 100 acres, you have an intense concentration of all kinds of sports and activities. You attended the cultural orientation right? I wouldn’t want to list all those possibilities again.

Try whatever activity or event catches your eye, whichever activity has that cute guy/gal you’d like to talk to, whichever activity your friends force you to take part in. Find something you’d love to do. But at the same time, don’t get lost in this world. There’s more to do.

I encourage thinking of college (and really life itself) as an RL environment. It’s worth trying things because even if you don’t get what you want, you gain information about how to get what you want. Your value function should intrinsically reward exploration, or at least your policy should mandate some. Something like Allen Naliath’s rejection challenge could also be a useful framing.

Kakkar also makes a point about lifelong learning: “Cultivate a lifelong learning habit. Wherever you go, whatever profession you choose after this, in order to grow you’d have to keep learning. You’d have to keep evolving.” In the “lifelong learning” spirit, I find this note on Andrej Karpathy’s Stanford page endearing almost:

I like to go through classes on Coursera and Udacity. I usually look for courses that are taught by very good instructor on topics I know relatively little about. Last year I decided to also finish Genetics and Evolution (statement of accomplishmnet) and Epigenetics (statement, + my rough notes).

You can almost certainly get full financial aid on Coursera courses, if you need it (this unlocks certification, but more importantly, graded assignments on certain courses). Of course, I think it’s probably more efficient to learn with LLMs now. That’s another college tip: explore the space of LLMs beyond free ChatGPT.

Kakkar also has a gameplan for a mediocre college:

Despite a high GPA, high confidence, and insane problem solving skills - if they don’t know you, they won’t recognize you or your skills.

Who are they? Anyone you want to impress: Investors, Employers, and Partners.

If they’ve filtered out your college, you can’t impress them by telling them what you did inside your college curriculum.

What you need to show is how you went beyond.

In practice, that translates into creating something tangible. An app, a website, a portfolio, a start up, or a game.

This Nikita Bier tweet is relevant:

The most important lesson for a young college kid trying to break into the world of consumer tech:

Life is a ladder so just try to get on the 1st rung. You can do this by building very small things that have a high probability of going viral—as a way to establish a name for yourself.

This means ruthlessly cut the scope of your ideas. No engineer, designer, or investor is going to give you much of their time when you have no track record. Small projects that can be built in a month can serve as a testament to your abilities.

I don’t think you need to make a “viral” product — people you want to work for can assess the merit of your work independent of its view count. But I agree that well-scoped, impressive projects can open vaunted doors.

Holden Karnofsky’s podcast episode on the importance of kicking ass:

Also, people who are good at things, people who kick ass, they get all these other random benefits. So one thing that happens is people who kick ass become connected to other people who kick ass. And that’s really, really valuable. It’s just a really big deal.

You often need to set time aside for kicking ass. See Mau’s “Many Undergrads Should Take Light Courseloads”:

If you want to go into research, research experience / track record and good grades (e.g. GPA of ~3.8+ for CS) on classes you took are major assets; having taken extra classes won’t be as noticed (and will burn time that, if spent differently, could better prepare you for graduate school / research).

80,000 Hours echoes this advice:

In fact, even in cases where good grades are useful to your future career, it could still be even more useful to pursue side projects, internships, running student societies etc. Going from acceptable to great grades often takes a lot of extra time, and this time can often be better spent on other priorities.

We’ve been discussing the importance of good grades, but in the US, there’s also the question of how many courses to take. While students sometimes take as many courses as they can, this leaves them with much less time for the valuable activities we mentioned earlier: getting good grades in the classes they do take, doing internships, doing side projects, meeting people, running student societies, etc. At the same time, taking more courses doesn’t add much to the ‘credential’ value of your degree. For these reasons, we think undergraduates can usually do better by replacing several of their would-be classes with other valuable activities.

Paul Graham’s “Undergraduation”:

But while you don’t literally need math for most kinds of hacking, in the sense of knowing 1001 tricks for differentiating formulas, math is very much worth studying for its own sake. It’s a valuable source of metaphors for almost any kind of work.

Graham’s “More Advice for Undergrads”:

“They may be trying to make you lift weights with your brain.” Indeed; I think pure mathematics makes excellent weightlifting.

I would say Leetcode is also great weightlifting for the brain. Sure, your day job won’t involve writing a linear-time algorithm to find a string’s longest palindromic substring (unless you’re a bioinformatician). The point of algorithmic problem-solving is mostly weightlifting for the brain.

On math, John Kerl’s tips for mathematical handwriting are great for taking notes during lectures. I find live TeXing unfathomable, but your mileage may vary: see SeniorMars’ guide.

Graham’s “A Student’s Guide to Startups”:

Most students don’t realize how rich they are in the scarcest ingredient in startups, co-founders.

Graham’s “How to Do Great Work”:

In most cases the recipe for doing great work is simply: work hard on excitingly ambitious projects, and something good will come of it. Instead of making a plan and then executing it, you just try to preserve certain invariants.

Graham’s “What You’ll Wish You’d Known” is directed at high schoolers but relevant for undergrads:

If I had to go through high school again, I’d treat it like a day job. I don’t mean that I’d slack in school. Working at something as a day job doesn’t mean doing it badly. It means not being defined by it. I mean I wouldn’t think of myself as a high school student, just as a musician with a day job as a waiter doesn’t think of himself as a waiter. And when I wasn’t working at my day job I’d start trying to do real work. …

You may be thinking, we have to do more than get good grades. We have to have extracurricular activities. But you know perfectly well how bogus most of these are. Collecting donations for a charity is an admirable thing to do, but it’s not hard. It’s not getting something done. What I mean by getting something done is learning how to write well, or how to program computers, or what life was really like in preindustrial societies, or how to draw the human face from life. This sort of thing rarely translates into a line item on a college application.

Ben Casnocha’s “Three Things to Unlearn from School” also touches on taking your work personally:

[B]egin working for yourself, and let the teachers be damned. But they won’t be – they’ll just be all the more approving because that kind of integrity can only command respect. After all, most of the work we devise is devised for students who are not working for themselves, so those that do surpass our expectations and teach us things that we’ve never thought of.

And so does Skunk Ledger’s “Escaping High School”, which is again relevant for undergrads despite the title:

Whatever your interests are, do the real version of them: get as close to the version that adults do as possible. Instead of submitting the story you wrote to your teacher or even to a contest, try submitting to some literary journals. Instead of writing a program for AP Computer Science, contribute to an open-source software project. Instead of studying for your biology test in preparation for pre-med, learn first aid and volunteer as a paramedic. Instead of playing in the school orchestra, busk on the street with your friends. Basically, make a bunch of stuff, no matter how bad it is — again, people’s expectations are low — and put it out there where people can see it.

Also relevant: Laura Deming’s advice for ambitious teenagers:

Say you are 14, and plan to go to college at 18. You have 4-5 years, an equivalent length of time to many graduate students or postdocs. Use this time to pick a significant project and make real progress.

(You’re likely not the target audience of Deming’s post — you’re probably in college and not 14. But then, this advice only applies more to you, since you know more than a 14-year-old does. “You have as much time in college as a PhD lasts” is a useful framing.)

On ambition, I found exurb1a’s video “You Will Never Do Anything Remarkable” inspiring when I was younger.

Nabeel Qureshi’s “Principles”:

“Aim for Chartres” (Christopher Alexander) — when doing something, aim to be the best there ever was at it. This compensates for your natural bias, which is to do something mediocre. You have to really aim to be as good as the greats.

unoptimal’s “Reflections on a year of college” collects lots of links to other college advice posts, and has some advice of its own:

If you’re not sure what to study (or what skills to learn), a useful heuristic is to learn stuff that maximizes your future optionality. These subjects generally seem to point towards the hard sciences, under the logic that it’s easier to pivot from a technical field to a general one, rather than the other way around. (And learning hard skills is useful.)

But see also Mihir Desai’s “The Trouble with Optionality”:

By emphasizing optionality, these students ignore the most important life lesson from finance: the pursuit of alpha. Alpha is the macho finance shorthand for an exemplary life. It is the excess return earned beyond the return required given risks assumed. It is finance nirvana.

But what do we know about alpha? In short, it is very hard to attain in a sustainable way and the only path to alpha is hard work and a disciplined dedication to a core set of beliefs. Given the ambiguity over the correct risk-adjusted benchmark, one never even knows if one has attained alpha. It is the golden ring just beyond your reach—and, one must enjoy the pursuit of alpha, given its fleeting and distant nature. Ultimately, finding a pursuit that can sustain that illusion of alpha is all we can ask for in a life’s work.

More college students should figure out what they want to do with their lives early.

Ben Reinhardt also recommends getting technical in “Precocious Young People Should Do Deep Technical Training”:

Technical training requires a level of immersion, undivided attention, and frankly – discomfort that you can only really devote when you’re young and relatively responsibility-free.

Finally, maybe I’m old school or wrong, but I hold the controversial opinion that deep technical subjects are just harder and more rigorous than others. If you can master them, you can master anything else if you put your mind to it.

Nate Soares’ “Half-assing it with everything you’ve got”:

If you’re trying to pass the class, then pass it with minimum effort. Anything else is wasted motion.

If you’re trying to ace the class, then ace it with minimum effort. Anything else is wasted motion.

If you’re trying to learn the material to the fullest, then mine the assignment for all its knowledge, and don’t fret about your grade. Anything else is wasted motion.

If you’re trying to do achieve some combination of good grades (for signalling purposes), respect (for social reasons), and knowledge (for various effects), then pinpoint the minimum quality target that gets a good grade, impresses the teacher, and allows you to learn the material, and hit that as efficiently as you can. Anything more is wasted motion.

Your quality target may be significantly left of F — if, say, you’ve already passed the class, and this assignment doesn’t matter. Your quality target may be significantly to the right of A — if, say, you’re there to learn the material, and grade inflation means that it’s much easier to produce an A-grade paper than it is to complete the assignment in the maximally informative way. But no matter what, your goals will induce a quality target.

Soares’ “Not because you ‘should’”:

Just stop doing things because you “should”.

As in, never let a “should” feel like a reason to do something. Only do things because they seem like the best thing to do after you’ve thought about it; never do things just because you “should.”

Alex Liu’s “Fury and Freedom: Four Years at Amherst College” is a short and vivid account of what doing something because you should looks like. Dan Wang’s “Violence and the Sacred: College as an incubator of Girardian terror” is the canonical reference on (not) doing things because people around you are doing them. Euan Ong’s “escaping flatland: career advice for CS undergrads” is also a good read on this topic: “planes of legibility” is a great phrase.

Soares’ “On Caring” is also great for thinking about what to do with your life:

The loss of a human life with all is joys and all its sorrows is tragic no matter what the cause, and the tragedy is not reduced simply because I was far away, or because I did not know of it, or because I did not know how to help, or because I was not personally responsible.

Knowing this, I care about every single individual on this planet. The problem is, my brain is simply incapable of taking the amount of caring I feel for a single person and scaling it up by a billion times. I lack the internal capacity to feel that much. My care-o-meter simply doesn’t go up that far.

If this line of thinking seems compelling, you should read the book Doing Good Better (free physical copy here) and check out effective altruism generally.

Alex Turner’s talk on “What do I wish I had known while at Grinnell?” given to students at his alma mater:

I’m going to state my advice now. It’s going to sound silly.

You should read a Harry Potter fanfiction called Harry Potter and the Methods of Rationality (HPMOR).

I’m serious. The intended benefits can be gained in other ways, but HPMOR is the best way I know of.

Trenton Bricken’s “Lessons from Undergrad”:

Learn frameworks for thinking.

A number of the neuroscience students I met were smart and hard working but seemed to graduate having spent a lot of time being taught the name of every brain region.

Along the recurring theme of there being finite time to learn, I wish I had better prioritized classes to teach me completely different mental models. Learning more about human rationality and cognitive biases; How to think probabilistically and in high dimensional spaces; Seeing the world through information theory and statistical mechanics; Learning about historical trends that have shaped societies and may do so again; Physics, biology and chemistry for the fundamental levers and constraints of our reality. I feel like all of these skills have outsized returns no matter the specific problem or textbook facts of the day. Feynman has made very similar arguments and this post is a great example of this sort of framework based learning centered around being able to understand any system.

Callum McDougall applies a similar sieve for deciding what math to study:

The three main reasons I usually saw people (including myself) choosing courses were (in some order):

How much they enjoyed the course

How good they were at the course

How directly useful the course would be (for future courses / university paths, or for careers)

When I looked back at which courses I found had been most helpful for me, I didn’t see much correlation with any of those three courses (except for possibly the first one). The common thread among courses that I remember today is that the overall flavour of the course (i.e. the kinds of problems it offered, and the kinds of reasoning it made me practice) was much more predictive of how useful I’d eventually find it.

Bricken also recommends you “[t]ake general education requirements you aren’t excited about at the very end of undergrad, not the very beginning.” I have written about this exact topic on my blog before.

Alexey Guzey’s “College Q&A”:

Q: should I double major?

No. Just take more fun classes. Nobody cares about double majors and you’ll only increase the number of boring classes you need to take.

Ross Rheingans-Yoo reframes the “what major(s) should I pick?” question:

The wisest advice I ever received about choosing your concentration at Harvard is “First write down all the courses you want to take. Then figure out which department will give you a degree for taking those courses.” For example, when I looked at the courses I wanted to take, I realized that they satisfied the requirements for either a CS concentration, or a joint CS+Math concentration. (Interestingly enough, CS and CS+Math have the same number of requirements, 12 courses.) So I declared CS+Math, and that was the end of it.

Zhengdong Wang echoes the “don’t double major” advice:

Don’t waste time taking classes you wouldn’t otherwise just because you’re “a few off” from getting another major. No one cares. I doubt that you honestly think the marginal benefit of the 24th class in just two departments is higher than the 1st class in new department. You only have 36 classes, ever. “A few” is a massive waste.

Wang’s whole post is worth reading. Here’s some advice that I would follow if I could (but cannot because I’m overloading courses ~every quarter so that I can graduate early, which starves me of choices):

Don’t take classes with section

… Even if you have the perfect TA, section is designed to fail. People don’t take lecture classes because they want discussion. I haven’t met a single person who likes section. No one in section wants to be there, including probably the TA. Sections are like bank runs. Even if you don’t think the economy is going to collapse, if everyone else starts taking their money out, you will too.

Juan David Campolargo’s “How to College: Advice, Mistakes, and Thoughts”:

Speaking up in the first class is scary. But speaking up in class 10 of week 2 is 100 times as scary.

The reason to do on the first day is to 1) lose the fear, and 2) gain momentum.

Henrik Karlsson’s “Childhoods of exceptional people”. He finds that prodigies grew up in exceptional milieus, roamed about and did self-directed learning, and took on “cognitive apprenticeships” — all examples of things you can do in college.

The Pmarca Guide to Career Planning is worth reading.

From the post “Opportunity”:

The world is a very malleable place. If you know what you want, and you go for it with maximum energy and drive and passion, the world will often reconfigure itself around you much more quickly and easily than you would think.

See this andi tweet:

pmarca’s “the world is surprisingly malleable” quote but it’s about asking out that girl you like

From Andreessen’s post on skills and education:

Seek to be a double/triple/quadruple threat. …

The fact is, this is even the secret formula to becoming a CEO. All successful CEO’s are like this. They are almost never the best product visionaries, or the best salespeople, or the best marketing people, or the best finance people, or even the best managers, but they are top 25% in some set of those skills, and then all of a sudden they’re qualified to actually run something important.

Andrew Conner agrees with this advice in “Career advice for young high-performers”:

So what skills should ambitious professionals focus on developing? The most important is being a generalist. Have a wide base of knowledge in many different fields. But often this means developing little depth. Don’t do this; actually develop expertise in multiple areas. You can have top 1% of knowledge and experience in many different domains.

Also worth reading on this: John Wentworth’s “Being the (Pareto) Best in the World”.

Also read the Pmarca Guide to Personal Productivity:

Let’s start with a bang: don’t keep a schedule.

He’s crazy, you say!

I’m totally serious. If you pull it off—and in many structured jobs, you simply can’t—this simple tip alone can make a huge difference in productivity. …

If you have at any point in your life lived a relatively structured existence—probably due to some kind of job with regular office hours, meetings, and the like—you will know that there is nothing more liberating than looking at your calendar and seeing nothing but free time for weeks ahead to work on the most important things in whatever order you want.

This also gives you the best odds of maximizing flow, which is a whole ‘nother topic but highly related.

Sigil Wen’s “Productivity”:

Write Essays to Articulate Clear Thinking

Paul Graham has lots of excellent essays on thinking by writing: “Writes and Write-Nots”, “Putting Ideas into Words”, and “The Need to Read”. Holden Karnofsky’s post “Learning By Writing” is also on this topic, and is one of the most important essays I’ve ever read.

On thinking in writing, I relate to Joan Didion, as quoted in Miller’s Book Review:

Thinking and writing form an inseparable pair in Didion’s process. “I’m terribly inarticulate,” she told Sarah Davidson in 1977. “A sentence does not occur to me as a whole thing unless I’m working.” Writing was how she thought. “Had I been blessed with even limited access to my own mind there would have been no reason to write,” she said in “Why I Write.” “I write entirely to find out what I’m thinking, what I’m looking at, what I see and what it means.”

She said something similar to Terry Gross in 2005 when discussing her book, The Year of Magical Thinking, on NPR’s Fresh Air. “It’s the way I process everything, by writing it down. I don’t actually process anything until I write it down, I mean, in terms of thinking, in terms of coming to terms with it.”

See Dynomight’s “Paper”:

The primary value of paper is to facilitate thinking.

For a huge percentage of tasks that involve thinking, getting some paper and writing / drawing / scribbling on it makes the task easier. I think most people agree with that. So why don’t we act on it? If paper came as a pill, everyone would take it. Paper, somehow, is underrated.