I agree. I usually us some sort of catalyst to help though. I like building things so I might watch something like this - http://www.youtube.com/watch?v=2c3q9K4cHzY - and then try to do it.
Or, for instance, watch MIT lectures from open courseware to try and learn how to code in Python.
Finding something you enjoy or want to learn should be the easy part. Then just try and do it using whatever resources you can muster. Pretty much everything we teach ourselves outside of mandated class work is considered self taught now isn't it?
I like this approach. Not because it gives you an easy start but because it forces you to look at a problem or business and think about how it could be done better.
You aren't really stealing ideas and copying people you are improving them. If you didn't the chance of success would be far slimmer. This is healthy capitalism. This process forces the other company to respond by improving and morphing. If they don't they risk becoming obsolete.
I have ideas all the time. A few good ones and a swarm of bad ones. Filtering them down is sometimes tough.
I believe you should work on the one you feel the most passionate about. Ideas do not need to be great to succeed. But, unless you have a fantastic game changer of an idea, you need to work hard and passionately to get it off the ground. The more you are vested in the project emotionally the more likely you are to always push through.
Hate cats but have a great idea for a new cat toy? Screw it. Love dogs but have an iffy idea for a new product (doggles anyone?)? Go for it.
As Edison said, "Genius is one percent inspiration, ninety-nine percent perspiration." Pick the project you are most willing to perspire over.
Or, for instance, watch MIT lectures from open courseware to try and learn how to code in Python.
Finding something you enjoy or want to learn should be the easy part. Then just try and do it using whatever resources you can muster. Pretty much everything we teach ourselves outside of mandated class work is considered self taught now isn't it?