Startup internships tend to be semester-long job interviews, which means that you'll be doing the same thing that the full-time employees are doing but with more supervision at first. This is a huge win for you, the intern, for two reasons:
1. Instead of doing menial grunt work you'll be doing real work for real users, which is a _way_ better learning experience. Finding out what people want isn't something that can be taught. You have to make things, put them in front of large numbers of people, and see how they react. You can get this experience by building a startup and going on the long march of getting traction. Or you can work for a startup that already has traction and observe their users. The latter is obviously way easier.
2. If you're good you'll probably get a job offer right away instead of an option to take a job after you graduate. If you're the sort of person that needs a piece of paper in order to feel like you're worth something this might not matter. But if you're in it for the work itself, you can save yourself two years of salary plus tuition.
In a related note to point 1, work for a consumer-facing startup if you can. The way consumers behave and the way business customers behave are completely different, and consumers are a lot more fun to work for.
In reply to 1 (having done internships at Facebook, Google and RIM), all of these large companies try very hard to avoid having interns do grunt work. These internships are as much about advertising the company to the intern (and the intern's peers when they return to school) as they are about getting work done. In every internship I've done with these big companies, I've had interesting projects to work on and was specifically shielded from doing some of the menial work that full timers had to do. I think some of them even had explicit policies about this.
I second this.. at Bloomberg we also craft specific projects for interns that can be completed realistically in the amount of time that they are there and are genuinely useful either internally or inside an external-facing product. Much of what is produced by interns is released into production.
That's not at all the same to working on the core product. Launching a new public-facing project isn't going to have the user base that the core product has.
I generally meant some kind of digestible deliverable within the context of the core product. You need to have a well defined goal that is achievable within the amount of time the person is there, otherwise it will not be a good experience for the intern or the company. So let me correct... most of what is produced by interns is part of the core product and released to production.
As a former Bloomberg intern, you're right – you try very hard to make sure interns are working on real projects that are deliverable over the course of the summer. That said, the projects are often completely isolated from what will have real impact on users on a day to day basis. This isn't a problem, per se, but I've had other internships where the work I was doing over the summer was virtually indistinguishable from that of a new full-time employee. I imagine this was largely because it was a much smaller company with a much smaller product.
I don't mean this as a criticism – Bloomberg tries very hard to make sure interns have a great experience and usually succeeds. But there is a difference and I, personally, preferred the other experience.
It also depends a lot on what team you wind up picking. Just speaking for the interns in our infrastructure team, we try to make them develop things that have an impact. You should let me know what group you were with so I can deliver the anonymous feedback :)
But what you'll be doing as an intern is not consistent across companies or even across teams. You might be QA, which is nothing like writing production code. You also might be assigned a project instead of working on the code base with the rest of the group. You might even be a sys admin or something similar.
If you're getting an internship, make sure you know what you're getting.
1. Instead of doing menial grunt work you'll be doing real work for real users, which is a _way_ better learning experience. Finding out what people want isn't something that can be taught. You have to make things, put them in front of large numbers of people, and see how they react. You can get this experience by building a startup and going on the long march of getting traction. Or you can work for a startup that already has traction and observe their users. The latter is obviously way easier.
2. If you're good you'll probably get a job offer right away instead of an option to take a job after you graduate. If you're the sort of person that needs a piece of paper in order to feel like you're worth something this might not matter. But if you're in it for the work itself, you can save yourself two years of salary plus tuition.
In a related note to point 1, work for a consumer-facing startup if you can. The way consumers behave and the way business customers behave are completely different, and consumers are a lot more fun to work for.