Nope, there are cliffs where the total number of bit-coins produced per day drop off a cliff. It's a silly idea designed to make early 'investors' lot's of money at the cost of making the entire system be less stable, and worth less money.
You realise there is no central control, right? The system can lag though, and if someone build a large rig and turned it on, that might have had an effect.
Why was it setup like that? Well it was decided rather than having a fixed rate of supply though time there would be a target number of total bitcoins and early miners would get a free pass: https://en.bitcoin.it/wiki/Controlled_Currency_Supply
A more rational setup would be a fixed rate of supply though time, so in 2750 the % increase in supply is a tiny fraction of overall supply but the absolute number would be the same as 2012. Instead they picked a fixed target number of bit-coins which eventually creates deflation as bitcoins can be lost which destroys them.
Every 210,000 blocks the mining reward is cut in half. At current rates, the mining reward will drop from 50 to 25 btc around December 2012. This is separate from the difficulty updates that occur every couple weeks.
Yea, and what protects the bitcoin network is not how much effort it took to create them in the past but how much computing power is being used to create them now. As the rate of creation drops the incentive to mine them also drops, eventually the number of miners drops off until it becomes cost effective to rent enough computing power to own the network and give yourself 100x the amount of bit-coins in existence. Cash out and well game over.
As I understand it, the goal of this feature is to drive miners to begin charging transaction fees to mint blocks rather than rely on the reward built into the protocol. The reward was designed as a carrot, but eventually it'll give way to transaction fees. If you study the protocol, transaction fees are built in; early on, however, they weren't used (i.e. spending BTC was free).
However, there is a more interesting question about transaction processing rates and competitiveness. VISA claims processing rates of 10000 transactions/second. BTC is at around ~700 transactions/second.
The problem boils down to this, suppose total bitcoins where worth 1 billion dollars, an average day exchanged 1/10th of that or 100 million dollars worth of transactions. Now suppose a 1% fee per transaction that's 1 million dollars worth of computing power per day or ~41k per hour (assuming mostly rational actors). Now suppose it costs 10x that or 410k per hour to take over the network. Well the 'pot' is worth 1 billion so that's a clear win and someone would do that.
