One theme across most of those stories is that the author and his team went out and built something before validating their ideas. Actually building a product is perhaps the HEAVIEST way to validate whether there's a market for the idea. Spend more time upfront validating the idea... make a fake version of the product, create some landing page that people can sign up for... anything to avoid going out and building before you know you're on to something.

It's okay to mess around with products in your spare time if it's just 'tinkering', but if you're trying to do the startup game for real, then lower investment upfront in ideas is critical to finding one that works.