Google will be the Microsoft of the decade.
Microsoft will be the IBM of the decade.
Microsoft will be the IBM.
Facebook will be gone in 5 years, just like MySpace. Half a dozen companies will rise and fall in replacing it. The web will churn even faster than it did before. Computing power is reaching a tipping point and I think we haven't even seen the tip of the iceberg yet in terms of the volatility of the industry.
Imagine 10 years from now where the computer power of facebook, twitter, or hulu today is the equivalent of a $20/mo. Linode account. Imagine a 1U pizza-box server with a terabyte of RAM, a petabyte of redundant SSD storage, and hundreds of processor cores. That sort of technology is coming, and as it arrives it's going to transform the industry.
Imagine a world where it's possible to scale out to not just millions but billions of customers practically overnight and at low cost. It might be possible to have a million transaction per hour service used by half the population of Earth run by a single-person startup that is only just barely ramen profitable. These levels of technology are coming, and they will absolutely be game changers. Web service companies will be able to grow at tremendous rates, become dominant overnight in newly created markets and then be eclipsed the next month by some yet smarter company.
Twitter and Facebook rose faster than Google, Google rose faster than Apple, and Apple rose faster than anyone thought possible. Not only will the next companies rise even faster than Twitter/Facebook but the cycle will be shorter, and the next companies after that will rise even faster in an even shorter cycle, etc, etc.
Anyone who thinks they can predict what the state of computing or the state of the web will be in 2020 is just fooling themselves. It will be different than today, that's about as much as we can know for sure.