As far as my experience with NLTK and machine learning in general goes: you'll need to spend a fair bit of time figuring out which features to extract.
I really like the Lenovo X201. Good build quality and extremely portable. You'll need to get a docking station if you want a DVD player, but other than that, it's a pretty nice machine.
Gadgets: Buy a u-lock (harder to cut) rather than a chain, wire or some other configuration. Buy a cycle computer (shows you your speed, distance for your ride, etc) - the feedback is great. You'll know when you're pedaling slower than usual and also can keep track of how far you've been riding. Last of all, buy a pump for the tires. Keeping your tires properly inflated helps a lot.
Contrary to what has been suggested before, I would buy the lightest bike you can in your budget. A lighter bike will reduce the agony when you don't want to push yourself; and when you do want to push yourself, just go faster.
Great feedback. I definitely am going to consider buying a cycle computer. Do you have any recommendations? I wonder if any services exist to transmit my stats? Say how many miles I do in a day or other riders. It could be a game.
You can also get hired at a research group in a university without a PhD (as a full-time developer for example). The pay is not competitive with industry, but you get to work on interesting problems. I do this at Penn State.
"Please don't water down my PhD ... by teaching me management skills - I'm not in business school.""
Most faculty manage groups of 5 - 20 people (their lab), a budget (their startup and other research funds) and have to deal with all the problems of running a team of people with varied skills, interest and motivation. Some faculty go on to be department heads or directors of institutes. In all these cases, management skills would be useful. Sure, you're still doing research, but you'll spend a good chunk of your time managing the lab.
"These are the people we can trust with tenure, because they would do research even if we didn't pay them at all"
It's this attitude that keeps the postdoc (and in some cases the junior faculty) salaries so low. Just because you love what you do, you shouldn't have to raise a family on $40,000/year
The higher you climb, the less hands-on research (including everything from planning to write-up to manual procedures) you do -- absolutely. It's staggering how much money is wasted in the process due to lack of management skills.
One solution would indeed be adding managerial skills to the PhD curriculum. However, most PhD will never reach such a "position of power." Hence, the more efficient way around this issue appears to be hiring of full-time lab managers.
Thanks for pointing this out. I agree with you that having managerial skills would be important to someone who manages people, and that many faculty in academia do need to manage people.
My point is not so much that nobody needs management skills, it's that not everybody needs management skills. Let me decide whether or not I will need those skills for myself. Perhaps I already have some management skills from an undergraduate education, or perhaps I simply have a natural talent for it. If I think that I need management skills, I will learn them - I have plenty of opportunities to take such courses or seminars, if I want. I'm an adult, you don't have to force me to do something because I might need it later, just tell me the facts and I will decide for myself.
But if I think that it would be a better use of my time to do something else, like collaborate with a researcher at another university on a new problem, let me decide that. I hope that whatever job I apply for decides whether or not to hire me based on my specific skill set.
I guess what I'm trying to say is this: a PhD is not (or should not be) a certificate declaring you qualified for a job. Not for an academic job, nor an industry job, nor any other job. It is a certificate declaring that I am very knowledgeable in a specific subject, and that people better and wiser than me believe that I can do meaningful research in the future.
If I want to show people that I have other job skills, I must show them some other way. A PhD only verifies the above, it doesn't verify that I am a good teacher, a good manager, a good communicator, a friendly guy, or that I am in any other way competent. If a job wants skills that I don't have, they shouldn't hire me.
For your other point, I totally agree that this attitude keeps the salaries low. Unfortunately I don't know what should be done about it. Either we can lie, and pretend that we only enjoy our research to the extent that it pays the bills, and otherwise do not enjoy it, or we can actually try to stop enjoying our work. Either seems silly. You can't hide what you love, and if other people try to take advantage of the fact, well, so be it. In fact, I can understand the counter-argument: just because work is important, doesn't mean we should pay people more than the bare minimum they will accept. It's not fair, and it isn't right, but it is rather true.
If you don't like the mostly dry quinoa that's so popular in the US, try it in a soup. That's how the Peruvians make it, and it's a world of difference!
At a number of places I've worked at, people did not want t-shirts. They preferred a free meal instead. Mentioning that a t-shirt lasts a lot longer than a meal had no sway.
As far as my experience with NLTK and machine learning in general goes: you'll need to spend a fair bit of time figuring out which features to extract.