There's not enough computing resources available to make it 100x the current rate.
At current difficulty, no. If difficulty falls to 1/4 the current level (like it was when we last dogpiled on), yes. Ok, the 1000 was a bit of hyperbole, but if all BTC miners switch for a day at 15k difficulty we could mine out all 2016 blocks in around 24 hours. Leaving namecoin with a 1.5 million difficulty, and several years of massively subsidized mining at a rate of a block per week until it adjusts back.
In reality something short of that will happen.
Come to think of it, if the US government wanted to kill bitcoin today that'd be the way to do it. Ask the NSA to switch their cracking lab to mining for a few hours over a weekend, rack the difficulty up to say 20-30 million and bitcoin is history. The fact that it hasn't been done proves to me that the US gov't, at least, is utterly unconcerned with bitcoin.
I agree with your principle but not the actual numbers. At a difficulty of 26,000 it would take ~42 seconds per block and ~2650GH/s (total network hash rate) to make the 24hour mark. But that would amount to 26,000 x 14 = 364,000 difficulty and NOT 1.5 million. Once again I think you exaggerated a bit on your calculations. But I must agree with all of you, once the difficulty of NMC adjusts downward by 50% watch some fireworks take place.