[karpathy]

It seems that a larger-than-I-expected fraction of the response to my post concentrates on a very small part of it, especially the part where I enumerate some considerations for thinking about whether a PhD might be a good fit for you.

By far the largest fraction of the post is concerned with tips/tricks for effectively navigating the PhD experience once you commit to going through it. I jokingly refer to it as a "survival guide" because (as I mention in the disclaimer paragraph) the experience is by no means a walk in the park.

[thr0waway1239]

Also do you mind modifying the title to say a Survival Guide to a C.S. PhD?

Absolutely everything the GP says is spot on, the C.S. PhD is really atypical amongst all the Ph.D. disciplines. E.g. the resource constraints that exist in many disciplines don't even exist in C.S. - it is almost completely the product of thought/ideas and ever cheapening computing power. I would suggest sending out your article to Ph.D. students in different disciplines, look at all the feedback, and then incorporate them if you are so inclined. The article might look quite different.

Edit: after reading the comment from pgbovine, I felt I made an unnecessary personal comment, which I omitted. I rewrote the sentence to keep the main point.

[pgbovine]

My unsolicited advice: Don't change a word.

At most, link to this HN thread so that readers can see different perspectives, but you wrote the article you wanted to write, not the one that anonymous online critics wanted you to write. If others want to write a response saying why you're misguided, by all means go ahead.

[roel_v]

I disagree. It's true that it's easier to get funding for CS but the career strategies described are exactly the same for most other fields. Many phd students just can't understand this post though; they think they're 'students' like they were in undergrad but it's whole different ballgame. The concept of the symbiotic relationship between adviser and student for example - most phd students just cannot see that, even after it's spelled out for them. Ito requires a certain mental maturity that most 23 year olds just do not have, it seems.

[thr0waway1239]

A friend of mine did a PhD in Biochemistry. Here is what I learnt, based on the many conversations I had with him while he was doing his PhD:

Freedom to choose the topic: Only to the extent that he could choose the advisor amongst a list of 3-4, which gets narrower after advisor selection

Ownership: I understand the sense in which the author wrote this, so I cannot disagree too much. But my friend would often explain how he could be most productive precisely by being a 'cog in the wheel' in the sense of how much cooperation was required from his lab mates for him to make any amount of reasonable progress

Exclusivity: The exclusivity is true in the sense it is described, but unlike in CS, did not lead to any major benefits (so sort of diminishes the leading statement about the appeal of the PhD). He went through a couple of post-docs, and then eventually landed what was a coveted position in his field - which still pays not very much

Status: This is the only part which I completely agree, and I actually respect non C.S. Ph.D.s all the more for their persistence because there is very rarely an escape hatch.

Personal freedom: Almost 100% not applicable. As I mentioned before, my friend could not make any progress without a ton of cooperation from lab mates, needed to be in the lab usually based on timings of other lab mates.

Maximizing future choice: Nearly every discipline other than CS would disagree with this. If you read stared's story, you get the distinct sense that his choices were maximized because he came into data science (i.e. he didn't feel that way while in his discipline)

Maximizing variance: This is immensely difficult for science Ph.D.s from what I understand, because the process from Bachelors to PhD, and often with a PostDoc or two on the way, is already too long for most of them and takes up their best years. So the statement "You’re young and there’s really no need to rush" is, well, a bit impractical.

I cannot comment on 'Personal Growth' and 'Expertise' - I don't know if you need a PhD for the former, and the latter is wonderful as long as the cost is not exorbitant (this exorbitant cost is common in other disciplines)

If the advice here is widely generalizable, then I would really like to see a few links to PhDs in, e.g. the physical sciences talking about their experiences with similar pleasure.

On the other hand, the stories I heard from my friends who did their Ph.D.s in other disciplines (Chemical Engineering, Mechanical Engg., Civil Engg., Chemistry, Physics to name some I remember) all had very similar patterns in their horror stories of the lack of resources and its impact on their journey.

So, is it at all possible that a CS Ph.D. who worked on an excellent topic (for which karpathy gets my kudos) in an internet-friendly, internet-visible, exploding field of work at a top institute might not paint anything close to the full picture?

True, that is the not the full article. But the best way to be safe is to never get into a fight - don't get into a PhD without understanding what you are signing up for. The section answering the question: "First, should you want to get a PhD?" is not well researched or widely sourced once you consider that Computer Science is not the only discipline in which people get their PhDs. Hence the suggestion to change the title to C.S. Ph.D.

[stared]

Yes, it is a small part of text, but one that may persuade someone into doing PhD, or give a false impression (e.g. the typical one, in which I used to believe: "follow your dreams in academia or get money at a dull job").

(BTW: I guess you know http://www.pgbovine.net/PhD-memoir.htm. Also from an uber-successful PhD student, but the full story, rather than a set of advice.)

It may be something about the field (growth, competition with industry). I think it was not a coincidence that out of many friends of mine who did their PhDs, only Wojciech Zaremba (now in OpenAI) had some non-trivial impact on the world.

I don't want to imply that even if everything works (topic, advisor, funding, the sense of meaning, the sense of progress, ...) it is any easy path. And I am really sure that even with your skills, work ethics (and luck) it was a challenge. Still, even if one field is rosy (DL or maybe CS in general), a typical PhD experience is hardly sth I would recommend blindly (vide links there: https://pinboard.in/search/u:pmigdal?query=academia+depressi...).

(On an unrelated note: thank you for "The Unreasonable Effectiveness of Recurrent Neural Networks", ConvNetJS and CS231N - they brought me into the deep learning world. :))