I'd focus on working on projects and rolling them out. See if people use them. If you can find a way to monetize, do it. Just build things to get experience. Don't try to start a start-up just cause.
I guess my behavior is the exact opposite: I won't start a project then try to monetize it, I prefer to look at what people are likely to spend money on and then build that. That seems more in line with what the OP is trying to do.
I can't tell you the countless times my college friend came up with a "great" idea. Even went as far as registering an LLC. I quickly dismiss most of his "great" ideas with 5 minutes of googling.
Starting a company means you need to invest ALL of your free time into building the company, product, acquiring customers, etc.
Most people register an LLC with their crappy idea and think they are running a company. Unless you are willing to drop out of school and invest 80 hours a week building this company, it's probably not worth it.
The reason PG suggests to build things that don't scale, is because getting experience building applications is far more valuable than trying to force a stupid idea into a business. Building applications that already exist just for fun, will open your mind. Once you start connecting different ideas together, organically, an idea worth becoming a business will come to mind.
You can sit and brainstorm ideas all day, but the best ideas, the ones that can turn into a business, they just come to you out of the ordinary.
The ones that are truly start-up worthy are those you personally believe in, have a considerable market opportunity, and the timing is right. Those are more rare, but if you make a concerted effort to keep thinking and re-thinking your ideas, you'll come up with one for sure.
Don't make a business to be successful, that's usually going to end in tears. Make a business to scratch an itch you have you can't ignore. Even if you fail you'll still feel like you've done something important, and if you succeed you'll be satisfied.
Obviously you'll push for the best possible outcome.
The alternative is like jumping out of an airplane with no parachute and hoping to hit a haystack. It's probably not going to happen. Most startups implode before they get any traction, it's a fact.