No. Just until 2013 or so. After that GPU mining will most likely be unprofitable. FPGA mining coupled with the reward halving would mean high efficiency is needed to mine. You won't get that from GPUs. Later on ASICs will make the game even harder.
I bet my GPUs will continue chugging along. Mostly because they basically run for free, as they are within the flat rate I pay for my bills. Lots of GPUs run on "free money" and will continue to run post-Jan 2013.
Profitability depends on BTC price and difficulty. It can perfectly well adjust to be profitable (say, go up to 12$ with 20% more difficulty than now) and predictions are hard at this point. If you are totally sure of your scenario, go and speculate heavily on it. You will be rich.
The key to the OP's question was something that took me a while to find out initially. The answer is that while you currently have a subsidy for solving a block, as you know it will eventually halve into nothingness.
However, just because the subsidy is no longer there, doesn't mean you can't solve more blocks. The solving of a block doesn't necessarily require generation coins to be paid out - indeed, you could claim fewer generated coins than are actually available, if you wished to.
Therefore, you can continue to solve blocks beyond the elimination of the subsidy, and the block solving is still necessary to allow transactions to be processed.
By that time, it is expected that transaction fees will be the norm, and will also be enough to make mining profitable still, which would be the incentive for mining and therefore keeping the network secure.
This paradigm switch is a test we still need to pass.
Personally I think it's a bad idea to subject the network to a traumatic test every 4 years. Speculation will go nuts every time we approach these milestones. It's going to be interesting to say the least. I'd rather have a flat production rate, which will have a similar result inflation-wise (approaching 0% growth gradually).
blockchain... erm ... bloatchain will be 1000 terabytes by that time, and bitcoin would be dead cause you'd have get 10 empty hard drives just to hold it all...
It's completely absurd to make predictions so far in the future. There are mechanisms to reduce space drastically, and also storage density per $ has never stopped doubling every few years since the 50s.