Ah, I see. First, no legitimate node would accept the block that has time in advance of more that max clock drift. The max clock drift is currently 2 hours. It came from original bitcoin client and for bitcoin network it is not a problem at all. As I mentioned, I am going to change that parameter, but it requires a hard fork, so I would like to collect all such hard changes together. Also no legitimate node would accept the block that is in the past relative to the GetMedianTimePast().
You could of course (and it seems you did ) make a tweaked wallet that would persistently submit the blocks with time 2 hours in advance. This would influence the difficulty of course, but will not stop the network from generating new PoS blocks as I assume that even before your dump you were not in possession of 51% of all coins. So even if you managed to bring the network's PoW difficulty to some high value that 'stopped' the current network's hashing nodes from producing PoW blocks and you started to create your own fork to override main chain you most likely would fail as the cumulative trust score of your fork would be less than the network's one: the block trust score is a function of the block's difficulty and a history of last 12 blocks. So, to achieve the better result you should have more resources (both the hashing power and the supply) than the rest of the network.

Taking to account the currently low network hashing power it is of course possible to make successful doublespend, but as I already said I am not going to pump the situation with the thin air. As the coin presently has very low value I do not see any reason to care much about this: when and if the graincoin get well valued then miners come.