A Survival Guide to a PhD (2016)
karpathy.github.io
karpathy.github.io
If people fail, it is mostly because they burn out. If they succeed, it is not unlikely that they will need to heal their burnout wounds anyway.
I am sure Karpathy's experience is different. But most people starting their PhDs are not Karpathy.
See also "The Lord of the Rings: an allegory of the PhD?" http://danny.oz.au/danny/humour/phd_lotr.html
Sure you may survive. But even if all goes well, you succeed, there will be a void in you after the quest.
Your supervisor may not understand this until it’s too late, and you may not have the ability to judge your adviser's ability to do so until you are committed.
The main problem is that you were raised in a school system where if you show up, study and do your assignments you are pretty much guaranteed to succeed sooner or later. A PhD is not like that.
But as you said, a PhD is quite different than all schooling before that. And that's good. A PhD is supposed to signify that you contributed new scientific value as judged by the expert international community, not just your teacher. Of course there are many wrinkles on this story like sloppy knee-jerk reviews etc.
But anything in life where you "just show up" and fulfill some explicit assignments tends not to be very valuable. If just showing up and doing what someone else decided for you is enough for a thing, that thing will lose value very soon. Similarly if you make sure almost everyone can do it, it won't have value, but will become a participation trophy.
But nothing in real life work like that. School is fake. You don't get a job just by showing up or having a diploma. Nobody will fall in love and start a relationship/family with you for showing up and fulfilling some list of criteria. Nobody will fund your startup or strike a business deal with your company because you showed up and did some assigned tasks.
In almost all aspects of life being proactive and exercising agency will get you much further than the teacher's pet mindset that school instills. And unfortunately rather than selecting for it, the PhD selects against such agency again because it's the safe option and people who are ready for an adventure usually dislike the academic environment. Not all of couse, I obviously don't mean every single person fits this. But in my experience this explains part of the mismatch in expectations and reality for the "I was a good student so a PhD felt natural" people. Not those come into the PhD with a well thought out plan, and knowing exactly why they want to pursue it, the upsides and downsides etc.
The "date-me doc" community might disagree:
* https://www.nytimes.com/2023/08/02/style/date-me-docs.html
* https://www.astralcodexten.com/p/in-defense-of-describable-d...
* https://www.astralcodexten.com/p/highlights-from-the-comment...
I'm not new to dating educated professionals, and it seems to be an unfortunate recurrence balancing being torn between wanting a partner and family and worrying that any misstep could jeopardize their career or cost them opportunity.
There exist a large number of very smart people who are not that career-obsessed (the traits that you need for a fast career are rather different from "highly smart").
I consider it to be likely that you value character traits that made the partner predestinied for a successful career, but don't like the fact that because of this, they are often obsessed about their career.
Including paramedic or trauma surgeon, where the explicit assignment is to stabilize the human being in front of you?
Or how about plumber, to stop water where it shouldn't be, making it come out faucets and go down drains where it should? Had furnace issues when we switched it on in October, the HVAC technicians made the magic box generate warm air: his explicit assignment was 'make the house warm' and he showed up and did that.
And in some sense we can make the PhD requirements explicit too: prove that you are able to provide valuable and novel contributions to the scientific community as recognized by them. But it's not box ticking. But so is becoming an especially appreciated and respected doctor in your city. It's not just don't kill too many patients and you're golden.
If you want a mediocre run of the mill PhD it's not super hard. It is hard, but not super hard if you have academic aptitude. You can do something incremental and follow the trends, and eventually some venue will publish it. Not many will read it though and you won't become a leading voice though. There are many incremental paper that help many average PhD students get their degrees.
There are plenty of other non academic fields where you have to work hard and long hours, such as finance or law firms. You can choose less demanding alternatives too but they are less prestigious. Of course finance pays better.
I think the problem is that people want to treat a scientific career as both a regular job and also as something of a pop star personal brand type of thing, and they only want the upside of both.
There are many problems with academia, power balance etc of course. It's not unlike problems for aspiring movie stars or in other somewhat isolated bubbles within society like in sports or the theater scene, or how kitchens work in high end restaurants etc etc. If you want to become an acclaimed top expert in a subfield, it's not easy. And humans are going to do their human things around such opportunities for status advancements.
You forgot about the humanities, but it's hard to blame you. So do most grant offices.
In practice these PhDs use the same label but have diverged a long time ago and their cultures, expectations and problems are quite different. As a STEM person, I also don't exactly understand what gets you a humanities PhD. I assume something like coming up with a new interpretation of a text, event or intellectual movement, connecting dots together in new ways that appears insightful to other scholars. But it's probably different in history, literature or other branches of the humanities as well. As I understand novelty is always part of it, as is significance in some way. You need to defend some kind of novel thought of yours, in a thesis.
It was not my case and bold of you to assume so. I had peer-reviewed publications before I even applied for PhD.
While I do know some people who expected PhD to be "more classes with more difficult assignments", the mast majority of PhDs I know had nothing to do with mentality you described.
I’m aware the experience will not be the same everywhere as PhD programs can be quite different, but many students that asked me if they should do a PhD are surprised when I tell them it’s not just up to the university or their advisors that they will get their degree and even more surprised when I tell them how early they have to start submitting their papers.
Even though I was well warned by my seniors, I admit that I also had this misconception to some degree.
1. You can hedge your bets by submitting your work to various conferences of various qualities (without going 3rd-tier, you can bet across 1st-tier and 2nd-tier)
2. You can spend time choosing the professor and the topic before going all in
3. You can seek advice and social interactions within your research group, departement and school
None of this is a silver bullet, but it compounds in the right direction.
To a point. If you stray too far from your personal interest, you increase the risk of burnout.
My advisor was well established, tenured prof with a number of students. I had to teach, but the effort was light. We taught large, basic courses that are boring for tenured profs. We usually requested the same 1-2 classes to teach and after the first round had all the materials (homework, quizzes, etc.) and could teach on autopilot. University gave us undergrad graders to grade assignments but I never used them since I wanted to see what my students wrote. Which is a testament that the load was light; if I was drowning I would use all free help I could.
But there was a cult of academia. "Get an academic job or you are a loser" mentality was prevalent. My advisor was disappointed, but OK when I decided to go into industry after PhD, but a friend's (Physics PhD from Harvard, CEO of a profitable startup now) advisor does not talk to him anymore because he did not stay in academia.
And I only realized long after finishing my PhD how incredibly lonely PhD path is. You live in your bubble for years, with minimal interactions outside a few other folks at the same grad school. Stipend was enough for basic living, but not much else. No good vacations, ski trips with friends, etc. And a few somewhat creepy characters that grow in this lifestyle. This is all surmountable, but the mental toughness required is certainly something to keep in mind. I did not have that mental toughness, but was an introvert, which helped a lot. But looking back I see that I also could have gone off the rails. My 2c.
It depends highly on the field. In history, sure. The point of getting a history PhD is to become a history professor, and you can't do that if you don't get the PhD, and meanwhile history PhDs don't meaningfully open up any other job prospects, so attempting and failing to get a PhD provides negative value.
In CS and many engineering disciplines, there is a long history of people dropping out of PhDs and landing in industry. The industry is therefore much more accustomed to, and therefore accommodating to, people taking this path. Whether it's a maximally efficient use of time is another question, but it's certainly not wasted effort.
But I do agree that it's stressful nonetheless because it still feels like a failure even if it is not actually in reality. I wrote about this when I put down my own PhD journey here [1]. In particular after the control replication (2017) paper, I very nearly quit out of academia entirely despite it being my biggest contribution to the field by far.
[1]: https://elliottslaughter.com/2024/02/legion-paper-history (written without any use of LLMs, for anyone who is wondering)
Do you mean that you’re using AI as a search engine for your local bibliography? I haven’t seen any AI plugins for Zotero.
On the zotero front there a bunch of AI plugins. But I've not used them. But yes the premise is that your can speak and ask your library questions. Some are set up differently though. Personally I can fire a paper into an llm and get a good idea of the content immediately and then ask questions about it. It's more interactive and allows me to get a better idea of it prior to reading it.
I only use it on a sentence or paragraph basis, otherwise it misses the point 90% of the time.
I would strongly advise against this use for the moment. The important part of reading a paper is not only to extract general rules, but to build your own internal model. Without it you cannot effectively do research. The main interesting points are often in the subtleties of the details deep in the paper.
Internal tought that come easily to mind when I read :
- 'oh they used that equation, but that could be also be interpreted totally differently, what happens if we change point of view, does it makes sense from this other perspective'
- 'I see they claim to achieve better results than sota, but actually, they compared with other methods that are not solving exactly the same problem, what shortcut or changes did they had to do to obtain a fair comparison, is it a fair comparison, can I trust those numbers? '
- 'oh, the authors didn't realize that they solved this other problem, or did they realize but there was a block somewhere preventing it?'
- 'I like this trick to achieve that result, but at the same time, it will prevent to solve a whole class of other problems, so their method will not work on those cases'
...
Also, notice that a paper IS a summary of multiple months/years of work, and researchers summarize it already to the maximum to stay within the page limit, by summarizing a summary you will always miss many things.
My experience with the bachelors was that despite my project being derailed by the bullshit around formatting the document, doing "research" by searching the library for peer reviewed papers that backed up my claims, etc, etc; I got a excellent mark. In short I set out to make something and due to the academic processes failed in making anything, but because I was able to critically reflect on it, I got a good mark. Waste of time, unless you were just are a good mark.
For my masters I know the project doesn't matter, I'm concentrating on the academic nonsense because that's where the marks are.
The waste of time would be for a professor to train you up to be a researcher before you’ve proven you are ready, hence the homework assignments.
I think the way to know if you want to be a researcher is more along the lines of: do you like finding the answers to questions no k e has thought to ask let alone answered? If so then it doesn’t really matter the training you’ve or the amount t of the field you’ve experienced, you can focus on that bit as your guiding force.
My suggestion is do a summer research project, and see if you enjoy it. If no-one will take you on, reflect on why that is (and to me that's a strong reason not to do it).
Add any paper you pick up to your tracking system before you read it. Make that part of your reading ritual.
Save the PDF right away, too. You may later lose access to the journal. Or, CiteULike (where I^Hyou uploaded all those articles) may go away.
It was the best/worst 4 years of my life. I studied overseas (uk), met my future wife and got a PhD that really wasn't useful for much to me. Fortunately it was under a scholarship.
1. It reduces the odds of missing a key reference in your papers and accelerates the write-up of the (often mandatory) Related Work section
2. It helps you maintain a mental map of the field as your research progresses
Personally I did the references at the end and didn't feel like I suffered from that decision, but the key references in my particular area were a relatively small and well-known set.
Personal freedom. As a PhD student you’re your own boss. Want to sleep in today? Sure. Want to skip a day and go on a vacation? Sure. All that matters is your final output and no one will force you to clock in from 9am to 5pm. Of course, some advisers might be more or less flexible about it. . .
* Your records are never good enough to completely replace your memory of what you did. The longer it takes, the more studies, readings, etc., that you will have to repeat.
* In physical and biological sciences, equipment breaks down, gets taken away, facilities get moved, etc. This stuff happens at a constant rate, and is a pure time cost.
* Technological progress gradually raises the bar for the minimum quality of some results, e.g., in computation. Even "theory" is highly computational these days.
There are also risks of career-ending accidents that can be treated as a constant risk per unit time:
* Your advisor dies, retires, gets promoted to administration, loses funding, changes jobs, gets embroiled in ethical / legal issues, etc.
* Some unexpected new result from another team or industry erases the relevance or novelty of your work.
* You get sick, have family crisis, etc.
* Burnout
Results are the wrong unit of measure. A better KPI is results per unit time. The people who look like they fucked around for 4 years then submitted a brilliant thesis were either working hard all along, or were just brilliant, which I certainly wasn't.
My then-fiancee and I were both grad students. We made a pact to meet at 7:00 every morning in the cafe across from the research building for coffee, to force both of us to stay on a work schedule.
What was the motivation? Honestly, I was too lazy to get a job and staying in academia for another 3+ years seemed amazing (probably not recommended, but it worked out OK for me).
What helped get me through it:
1) Doing something I genuinely enjoyed - I approached the Computer Vision professor who gave me some ideas. I super enjoy writing code, and the idea of processing gigabytes of video to produce answers seemed cool. I treated it as a super difficult programming project.
2) Breaking my leg - Just before starting, I broke my leg badly. And that meant working from home with a weekly visit from the professor with a stack of reading papers. That time spent understanding state of the art was super useful.
3) Funding - At some point, DARPA gave enough money for me not to worry about funding, so I never had to work a job or get distracted.
4) Marriage - The final straight of writing a thesis was tough and I was super lucky to have a supportive wife who pushed me to get-shit-done.
This is actually how I view academia. "Couldn't get a job"
It really lowered the prestige of a PhD for me. Heck, if I think through my PhD friends... none of them were A students. They were all C-tier.
Academia is a very particular dynamic very difficult to find elsewhere, and some people dig it. You can watch some people finding the same dynamic at Google for example, where they are allowed and encouraged to fiddle around and keep publishing (e.g. the Attention paper, why would Google allow such publication?). Such dynamics are explored in Terence Kealy book "The economic laws of scientific research".
It can be different in other fields an in lower tier colleges.
Something I didn't see in the article:
Depending on your field, it can be extremely hard to get tenure. Unless you are a genius and are seeing signs you are well on your way to getting professorship and tenure (publishing good and important papers, really good at communication, checking all the other boxes), you'd better consider an off-ramp as early as possible.
In some fields, 100 people compete for 1 open positions, and that's rough. Having been involved in such a hiring process, I know it's extremely challenging.
I was smart enough to realize I'd never be a good researcher relatively early during my PhD and started preparing for job interviews. Sadly, I see too many people still having unrealistic dreams about being a professor late in their PhD. They even do postdoc and do that for many years until they finally discover they are not hireable. Good news is that they often find an industry job quickly after switching.
I could see that clearly -- their research, originality, communication and understanding of the field is just not there, and one doesn't become a professor without being completely in control of the direction of their academic research.
But it's hard to tell people "you are just not good enough for this". In most situations, you'd be considered unsupportive. However, in this case, it's the best thing you can say to a struggling PhD/postdoc.
(Many professors are completely incapable of advising on students' career. They often find it surprising that a student wants to go into industry. They hold completely incorrect assumptions like a PhD student just goes into a management role as soon as they graduate. Rarely a professor tells you that you should not pursue tenure.)
What are some things that made you realize that? Im considering going into research and certainly have some doubts
* I had difficulty coming up with my own ideas. A good PhD student should develop a good sense of direction halfway through their research PhD, and by the end they should be able to conduct their own research independently. I was nothing like that.
* I attended talks from fellow students that were invited to the department and went to conferences. I know what good researchers are like -- they do good work, publish papers relatively early in their PhD, know everything about the field and what are they doing, have confidence, communicate with others well and have great delivery in presentation.
-- * This might be the one that is most compelling. You could tell who is going to be a professor and who is not. (As I am typing this, I look up the guy I have in mind. He just got assistant professorship last year. I am not all surprised.)
Granted, I am far from a model PhD student. My research is mediocre. But I have seen people who deliver even worse research and have even less chance of ever getting tenure spending more time in academia. So I hope my words can be helpful.
All that said, I would say you don't need to self doubt. If you are smart and constantly think about your career, and if you go to conferences, you'll know whether this is for you and what to do.
I feel like this particular advice applies to a very small subset of people. If I’d had professors telling me that I certainly would have considered doing a PhD!
How To Get Rich?
Step 1: be rich
Step 2: you're now rich
Karpathy is an exceptional person, maybe not as much as Ronaldo in football but taking advice from him similarly won't be guaranteed to work. You can't have guarantees in such things.
In truth, the more literal but not fully literal thing that happens regarding surviving a PhD is that you try to publish a paper in a top venue but after several rejections you publish in a lower tier one, then you do two more followup in similarly second tier but not terrible venues and you get a "magna cum laude" or perhaps a "cum laude" once you reach 5 years and the prof wants to avoid the embarrassment of not having graduated you.
Of course many people don't come into the PhD with such plans, they expect a summa cum laude and papers in top venues and talk invitations and so on, since they've always been a top student so far.
(I tried doing a PhD while working full time, and quit the idea after 3 years.)
When things went smooth (accepted paper) I tried to roll that motivation forward as much as possible.
2 years was during COVID so I never really had to travel for classes which really saved time.
I did most of the coding and ideation on either paternity leave or holidays. Evenings were often running smaller experiments or eventually endless iteration of a paper.
My wife was great and helped me focus usually 8-11pm most nights and some weekends 1-4pm.
We don't live nearby any family and friends were not much help.
― Richard P. FeynmanSounds strikingly similar to early-stage startup lifestyle.
(This is not to say you shouldn't do it. Just get info and advice from a less biased source).
Most people I know don’t use their actual skills anymore, but all of them shrug off whatever you throw at them at work without blinking.
You do what your priest/advisor tells you. You honor the priest. You do the latex ritual.
Did it actually do anything? Ah in 8 years someone is going to replicate your study, it wont work, but too late! You got a PhD!
Good luck to anyone starting out.
Perhaps, but the mindset of delegating your intellectual advancement to others really is not compatible with being in academia, let alone getting a PhD.
I'm certainly interested in learning and exploring and finding out stuff, but my experience so far is that the academic processes stand in the way of this.
Let directional universities pick first and Ivies (and other prestigious universities) pick last.
> The minimum salary for the 2025 NFL season was $840,000
Raise the minimum salary of a Ph.D. student to that level and we have a deal. (The pocket salary, not $830,000 self pay to the university and $10,000 for the pocket of the student.)
Also, the work in NFL is more standardized, all teams play the same games per week, have a similar amount of training time, ...
In a Ph.D. the topic depends a lot of the advisor. It would be like mixing all the sports in the same bag, and for a weird reason the Waterpolo team from Alaska can pick you that are an expert in Tenis.
Which is quite ironic, considering who wrote the article.
A PhD teaches you how to think, how to learn, and how to question the world. That's a vital set of skills no matter what tool exists.
I think there are many answers to this, not the least of which is that AI can’t really do it instead.
If that comes to pass, I guess there won’t be any economic cost to having done my PhD because the entire economy will be AI driven and we’ll hopefully just be their happy pets.
If that doesn’t come to pass, and AIs just remain good at summarizing and remixing ideas, I guess people with experience generating research will still be useful.
I liked my job at the university - independent of the final PhD. I enjoyed what I was doing. Most of the time I also enjoyed writing my dissertation, since I was given the opportunity to write about my stuff. And mostly I could write it in a way how I felt things are supposed to be explained.
Why go to the gym if you don't need physical strength? One needs to do something to not degenerate into a miserable state.
I find it very fulfilling to do a PhD and did so myself. More people should. What I mean is that I'm expecting the general view on it to evolve as described.