How far are your projections and where do you get quality inputs

Hi all, I have 2 questions:

1. Most books and resources I've read said that projecting earnings and financials more than 2 years away is not ideal, mainly because there is less basis to really judge cash flows so far away. My question is that I often only see earning estimates from reports 1-2 years (or even just a few quarters) away anyway, but does that mean you only project 2 years away in your model? Most of the models I've seen use 5 years (in M&A) but how would you make assumptions for yrs 3-5?

2. Going on with assumptions, one factor many analysts have cited that differentiate research is the quality of your idea (i.e. what is the story behind the numbers, the backdrop and angle to the research instead of the price target). In coming up with ideas and contexts in which to frame your research one may read broadly. In your work, do you have any interesting resources or sources you go to for inspiration, ideas or just different views?

Thanks!