This advice makes sense to me, and I would personally never want to open something that's a feature for someone else. However, there are startups in this space that get angel/vc funded all the time (Rapportive).

I guess I see the following scenarios could make this work:

1. Build a 'feature' that is complex enough that the other company will see value in just acquiring you rather than eliminating you.

2. Work for early mega-adoption so that you can pivot to larger scope/better product based on feedback