The advice I would give on a mentorship call

This is the post I wish someone had sent me as a freshman in college. I get a lot of mentorship emails. I probably call with one in ten of them somewhat randomly based on how much time and energy I have that given week, and I ignore the rest -- and I feel very bad about it. This post is some advice that I would give to those seekers.

Does it matter what I study in college? Yes and no. I studied Political and Social Thought, which is neither too closely connected to what I currently do nor is it a particularly “practical” subject in general, and I did just fine, so you really can make ~anything work. My major did teach me to write well and read difficult things and analyze them and discuss ideas with people, so I did gain some useful skills from it. But I'm also not a major relativist. I do think some majors are just better than other majors in that they are more rigorous or more well founded, come from a richer tradition, or teach you more transferable skills. For instance, physics just is better than media studies, math just is better than “american studies”, but say your passion is philosophy instead of mathematics you can make philosophy work, in fact philosophy can be quite good depending on the kind. I actually have a soft spot for the classics and think that they deserve a spot high up on the list of “best majors,” though you have to make a slightly more complicated and explicit argument for why it shows that you're smart and worth hiring.

How should you spend your time in University? You should do well in your classes and get great if not perfect grades -- most schools are not that hard and most have insane grade inflation. Also, if you are at university and getting mediocre grades people will wonder why in the world you are still there as opposed to doing something else you care more about, which is another valid option. 

Actually try. If you:

• Do all the homework

• Do the reading

• Study for tests

• Sit in the front row

• Actually pay attention during lectures and don't multi-task

• Ask questions

• Go to the professor’s office hours

You will be so much farther ahead than most students. 

To the extent you can, take classes because you hear the professors are fantastic. You may have to take some classes for a major even if the professors are only okay, but wherever possible ask students who the best professors are, and even if they teach kind of random classes, you will often get a ton of value from a particularly good professor. 

But don’t spend all your time on classes. Unless academic learning is your top passion, you should not spend all or ideally even most of your time on schoolwork. If you feel you need to do so in order to get great grades, consider whether you are in the right major or you have learned the most effective school habits. Because most people will not be trying, you can likely get great grades by doing what I mentioned above. You can learn to become a faster reader and writer, and it's ok to skip the occasional class especially if it is not one with a fantastic teacher, as long as you're studying for the test and doing the homework and paying attention to whether they dock points for attendance. 

Trying to maintain great grades while also not spending all your time on school is great practice for the rest of your life, where you will have to pursue excellence along multiple domains at once. 

Value your free time. While it may not feel like it, you will have some of your most free time during university. You likely don't have kids, you likely don't have a full-time job and a full-on life, you may live in a dormitory. In general your life will be simpler than it will be at other points. It also will be one of the few protected times for learning and so there is an argument that you should spend it doing that, but if you want to fill it with other stuff--

Do not fill your time with school clubs. Get out into the outside world. Try very very hard to volunteer in the outside world, to work, to intern. You will learn more and more relevant things by doing this. School clubs are a miniature, easier version of other activities -- go do the real versions to the extent you can. You’re ready, and if you aren’t, you’ll quickly become so by doing it.

How to get people to want you as an intern. The issue with interns is that they are usually more work than benefit. Any employee takes some time to ramp up, so when you get an intern for a few months most of that is ramping time and then they leave when they start to be useful. Interns require more instruction than slightly older new employees because they come in with less experience and life context. Even smart ones will still be missing basic tacit knowledge that comes from a few years in the workforce. And interns require the work of scoping and explaining projects to them, which can be painful to impossible for truly busy people. 

Ok, so given that most people really do not want interns, how can you increase your chances they still want you? 

• Be able to point to relevant work that you have completed -- eg. look at this report I co-wrote, look at this conference I planned. The more tangible work you can point to, the more reassured they will feel that you can do useful things. (You might say: ok, but the internship is me trying to get these things, but you can often co-write with professors or plan conferences / events at your school -- these are acceptable school based activities, particularly if you can do them in your first or second year). 

• Come in with ideas on how you are going to gain context quickly, eg ask if they can add you to the slack in the spring or whether you can come by the office a few times ahead of your start date and shadow meetings. You must promise that you will truly just observe, with the goal of coming in with more context so you can be useful faster. 

• Highlight that you are a quick learner and very coachable (this has to be true!)

• Highlight that you can make up for your lack of experience by eg. working longer than older employees who have families, being very happy to do unglamorous or annoying work/no task will be below you (this also has to be true!)

• In all of your interactions, show that you are professional, polished, and concise. They do not want essay-length emails from you. They want swift, clean, concise replies. 

• Tell them that you will be no worse than Claude because you’ll ask all your questions to Claude first. I would find this charming because it showed 1) they know how to leverage tools 2) they value my time 3) they actually wouldn’t be worse than a top of the line LLM which honestly is a fine intern baseline 

• They may or may not pay you. Payment is often less of a concern to them than you actually being useful, but it may be for nonprofits or very young startups. If they don’t pay you, you can often apply for grants from your school or foundations to cover your work. 

• Then, if they take you, be so good that they try to convince you not to go back to school in the fall! 

• Finally, recognize when some places are a lost cause. For instance, my current company, Anthropic, doesn’t do internships because they would be an information security nightmare, so ideally pick a place that is not aspiring to SCIF grade security. 

Cold emails. As someone who gets a lot of cold emails and doesn't reply to most of them, it might be odd for me to tell you to do cold emails. I do actually read all of my cold emails. The ideal cold email is sweet, specific, and shows that the person values my time. They explain who they are and why I should care about them, what they want from me and why me specifically, and ideally they have a doable ask. Maybe they have a question that they want answered in that email, or maybe they have a specific reason they want a short call. It's probably good for these emails to be somewhat flattering to the person. And they should never be essay-length. I have sent many cold emails. I have not gotten responses to all of them, but I have gotten responses to a few that turned out to be helpful or at least deeply encouraging. Cold emails are a good way to search out internships, though probably your first ask should not be an internship. It should be something else that is more manageable.

Getting your first job. A lot of this will repeat from the internship section. 

• Be able to point to relevant work you have done. This could be writing you've done in your spare time, open source projects if you're technical. The more you have a repository of legible work, the less of a risky bet you'll seem to your future employer.

• Study the business well. This should be basic, but not all folks do it. You should know a fair amount about the company when you interview. You should have read all their content, news articles about them, etc. You should have done your research on both the company and the interviewer. 

• The next more impressive level of that is identifying things the organization needs and being able to speak to those problems and solutions. It is quite impressive when a candidate can say “I’ve been thinking about the stage that [company name] is at and it seems like it might need X, Y, Z. If I got the role, I would love my first projects to be A,B,C that would address those.” Or “If I imagine a 99th percentile first year in this role I would have done [list of things that ameliorate concrete problems the organization has]”

• Prepare for your interview but don’t seem like you are reading a script. Have some talking points. 

• Send a thank you note to the interviewer and be specific about the interview and the role and company. Send it literally as soon as you get off the call -- small boost, but so easy you may as well do it. 

Being good at your first job 

• Do what you say you are going to do by the time you said you would do it (or ideally earlier). If you can't do it by that time, tell the involved party. Iron-clad reliability is just incredibly valuable. 

• Watch other people closely, particularly those that have skills that you don't yet have and want to acquire. You can learn so much through osmosis and imitation. Watch how they talk in meetings. Watch how they spend their time. Read the documents and messages they write carefully. Try to create a mental model of them. Pretty soon you will become more like them. 

• Be your own boss in the sense of looking around and seeing what needs to get done, and then trying to do it. Try to scope and find your own projects. Obviously check with folks if you've chosen useful things and how valuable they are. The more of a self-starter and self-manager you can be, the better. Try identifying ahead of your boss what they want you to work on and getting a start on it. You want to think "what would make their lives easier or what would make the organization's goals happen", and start on it. The more you can see your role as not just literally what is assigned to you but what needs to be done that you could plausibly do, you will become far more valuable. 

I received lots of helpful mentorship, and I owe a lot to the folks that answered my emails, met with me, and took a chance on letting me work with them. One mentorship tip that I intuitively got in college but temporarily forgot afterwards is that the most helpful mentors are usually folks who are only a few years ahead of you, because they experienced an environment that is much closer to your own and they remember it better. Yes, someone who is a decade ahead might be in an objectively more impressive place, or someone who is 4 decades ahead might have lots of time and wisdom for you in their retirement, but often you do want to try to find mentors nearer to your stage. 

Finally, thank your mentors!