One rule I've learned in writing about the future is to never say the word never.
Never as in, "You'll never have a Mac running on Intel chips." Never as in, "You can never be replaced."
This is especially true when covering open source. Never say never.
Expansive predictions just don't work in this space. So it's important that we get our heads around what open source is and isn't.
Open source is a business process. It's not Linux. It's not a specific license, like the GPL. It's not about any special piece of IT infrastructure. It's a business process, as distinct from a business model.
Open source lets a lot of people share development costs. As such it is a lower-cost way of doing business. Extracting that cost is always a challenge because you lose old business models in the process.
And I think business models remain the chief open source challenge. (I think Paul Murphy would agree but I will never say that for certain.) How do you pay for new development, for marketing, and for corporate infrastructure in an open source world? How do you make a profit?
There are many people meeting these challenges, and many who are failing at it. Never predict which will be which. Let the market do that. Or as Milo Mindebinder said in Catch-22, think of it as evolution in action.