I really liked "To test this you need to get real bullets in your gun as fast as possible. This means talking to potential customers and getting their feedback, implementing simple prototypes and measuring their performance, etc. Put your product and product ideas through the most rigorous process, using a combination of qualitative and quantitative feedback.
Then, using the data you gathered from your measurements and tests, iterate. It might turn out that nobody was willing to pay more than $20/month for your a simplified version of your product. This data lets you form new hypotheses."
Iterating is so important. Chances are you are going to always be wrong, so you always have to keep iterating. Especially in the early going.
I don't agree with RMS' zealousness either, but I thank him for all he's done. Sometimes you can't separate the end results from the methods. I don't think he would have been as effective and influential on the world had he been more reasonable in his approach.