Dear brilliant students: Please consider not doing a PhD.

I did my PhD in ECE specifically computer vision and had a decent time, but pretty much everyone I encountered during their PhD in biology had some horror story to tell.

I am approaching my final semester as a CS undergraduate and, to be totally honest, I am more than a little nervous about my decision to go down the PhD route. I've got a decent application, so I am less worried about getting into a program and more worried about life in it. I appreciate the experiences others have shared here, and I'd like to relate some of my thoughts and concerns because am curious to know if (how) I'm being naive. So, while my post here is mostly aimed at veterans who might dispel any illusions I hold, I would value any input.

I've applied to CS PhD programs and intend to do research in AI/Machine Learning. I am fascinated with the field, and though I have only a little undergrad experience with it, I've loved it all so far.

One of my main goals in getting a PhD is to improve my future career prospects, whether in academia or industry. I am ambivalent about going into academia for a post-doc career, as I think a startup or even industry would make better fits for me personally. I've noticed most industry jobs I'm interested in require graduate education, and I expect the PhD will likely over-qualify me for most startups. But more than all that, I want to make use of my mind (I know, I know: its that tired cliche about changing the world...) Though I enjoy coding, and I excel in my classwork and projects, I'm not a programming genius (which, perhaps regrettably, is what some people seem to expect from a top CS undergrad). I have had a taste of what it is like being a professional software developer, which is what my friends are prepping to do straight out of college, and I know it is not for me.

One qualm I have is that, as an undergraduate, I've had only one internship relevant to AI/ML and only a taste of research (no publications, some lab work), so the whole funding->research->publish sequence is still a little unclear to me (although reading blogs like Philip Guo's helped demystify the whole graduate experience). Another worry is that I do not want to put my life on hold. I can tolerate living on little funds, but I can't accept being a slave to my research: I want to have relationships and interests outside of the lab. To this end, I plan to keep my act together, work smart and effectively, and finish my PhD on time.

Does this seem like an unreasonable plan? Will I be facing jobless doldrums by having a PhD? What more can I do to be prepared for graduate school? Is there anything major I'm overlooking?

Also, thanks to HN in general for contributing - this has been a great thread.

EDIT: I neglected to mention I'm in the US.

You don't. A PhD is a full-time job, a 40-60 hour/week one.

I did the part-time Master's. That was a huge mistake. You will not get the education, the time to research, the networking, the collaboration, or the presence to do the job right.

A PhD done wrong is a waste of your time and the university's time.

The OP didn't really go into the "malice" part of it. FYI if you do good enough work your supervisor might be tempted to appropriate it. The process for doing so might involve painting you as incompetent.

It's not pleasant.

Quanics is correct. I would never recommend paying your own way at a graduate program--go where you are funded. A graduate stipend will net to about $20k/year (there are some amazing fellowships that will fund $50k+ but they are very, very competitive). In a "real job" you'll often be making much more than $20k, as I was when I was a higher ed IT professional (and could have qualified for free tuition as an institutional employee). PhD programs are rated by several factors including graduation rate and time to completion. It is obvious to me now that the people who are choosing the students to fill the very limited number of graduate student seats in any given year will choose those who are committing to work full time on graduating.

Many people do complain about academia being hostile to family life (see http://100rsns.blogspot.com/2010/09/15-marriage-and-family-u... for a particularly negative view). However, many of my colleagues had children during graduate school. They took time off classes but then used their "down time" (e.g. child napping) to focus on writing and came out ahead. I got married immediately before entering grad school and had my first child as I was writing my dissertation. The timing worked really well for me. I was lucky that my wife could support me through graduate school (on top of my stipend).