Having just got out, some things that I wish I had known when I started up:
Take Risks. Lots of them. College is one of the few times in your life when you're old enough to do cool risky things, but in a position where short of killing someone, everything is fixable. Most things that you do wrong (aside from plagiarism, which most places are big about going after) can be fixed, so take advantage of this to do some cool things.
Use your student status. Most modes of transport have steep student discounts. Your campus rec program will have classes that cost 30% (or less) than what they'd cost in real life. I got into martial arts because it got me into shape, introduced me to people outside my program, and gave me a way to burn off steam. You don't have to do that, but find a cheap class and take it.
Socialize outside your program. You make make fun of people in a lesser program, but they're interesting as well - just in other things. They provide some perspective on what you're doing, keep you sane and interested, and can offer great ideas without knowing it.
Socialize within your program. You may find these people boring, but you're all working on the same assignments and all having the same issues - so band together. They'll also be useful for networking in the future, and may have old exams you can study off of.
Find people you work well with. If you find someone (like a lab partner) you work well with, keep them - plan to take the same classes. I had a guy who I worked amazingly with in a lab, who was my lab partner for three years (across six classes). Never saw him socially, know nothing about him personally, but when we got into a lab we got shit done and fast, and knew how each other worked. This relationship is invaluable.
Get into student politics, lightly. I'm not saying run for class president, but there's almost certainly a Senate (or whatever your school's governing body is) committee on IT that takes student representatives. If you ask to be on this, you'll get a spot - since no one else will want it. You'll meet some interesting people in student politics (who will be good to know later), and you'll meet some administrators and department heads, who are also good to know - especially for recommendation letters for grad school. If your school has a technology transfer committee, get on that (it's full of interesting industry reps and people with great ideas).
Find a prof and get friendly. Every department has a few professors who are regarded as "crazy". They're the ones who have classes that people avoid because they're incredibly hard markers, and who teach a few esoteric subjects. These people fall into two categories - profs who mark hard because they're assholes, and profs who mark hard because they're demanding. Find one of the latter, take all their classes (which should be small because of their rep), and go to their office hours. They're likely very smart, and will provide valuable insight (and refereces) for your future. Get friendly with them.
Go to office hours and tutorials. Office hours are one-on-one with the guy who creates your exams, or at least superivses their creation. Tutorials are where that guy (or a TA he's instructed) goes over problems they think are important, which often are very similar to ones you'll be examined on. If the prof thinks that a particular problem type is important, then you can be sure that it will come back again - and it might even be important in the long run.
Don't sweat your grades. If you are sure you want to go to grad school, then sweat the shit out of them, and use your undergrad to form relationships with profs. If you don't, know that after your first job, no one will care about your grades. If you do good work outside of class and get soem good summer jobs, your first job won't even care about them.
Search for resume-builders. Ever school has a few societies that you can join for minimal effort that sound really good on paper, or classes that sound good but teach you nothing. These are worth it for the future. Also good is joining things like IEEE and ACM.