The difficulty is increasing by 30% per two weeks recently. Let's half that, 15%/wk growth. A 3MH/s PMB will only ever pay you out ~0.01 BTC.
The 'trick' here is that PAJKA isn't a 3mhs bond, it is a 15mhs bond. People buying now bets that BFL will deliver; simple as that.
Further, a sustained 30% per month is impossible for a few reasons.
1) Technology doesn't move that fast. Even if ASICs are way behind CPU chips for size, the Swedes are already working on 28nm ASICs. You can't go much below that currently, meaning distribution and commoditization must take over.
2) Increase in difficulty must long-term be funded by profitability; if/when difficulty increases beyond profitability, difficulty will not increase further.
3) Mining must compete with holding coins; this is possible only as long as Bitcoin/USD remains steady or falls.
Based on this, sure, we'll see a massive rise in difficulty, maybe we hit 1PH/s this year, maybe we'll have 10 PH/s next year. However, at that time or some point, buying ASICs won't make sense anymore because of 1, 2, and 3. New technology won't keep up, profitability will make it a loss to buy more hardware, and Bitcoins may have fallen further (in which case buying hardware won't beat buying coins, should the price come back up again).
It's very dangerous to extrapolate the future based on a short period of data, and by short here, I mean this year. For example, if we take the previous four days, the hashrate has dropped from 138 to 131 TH/s, meaning if we extrapolate that, mining will stop in roughly two months.
On the other hand, if we look at the rate from February to now, it's around 700% increase (over four months) indicating that we'll hit 6,5PH/s next February and 107,884 PH/s by next block halving, or around 300 million Jupiters (which is the most powerful scheduled miner today) or roughly 10 billion block erupter blades.
Even if someone comes up with 14nm ASICMiners this summer, and assuming those will be twice as powerful, we're still looking at every 1 in every 60 persons on the planet shelling out $7K for mining Bitcoins...
I'm sure KnCMiner wouldn't mind, but hey, I'm not sure it's very realistic.
PMBs lose value in massive difficulty rises, sure. So does mining equipment and it certainly doesn't seem like it's slowing down willingness to invest.
However, like I've mentioned before, PAJKA has been operating for over a year during which time it has paid out the principal plus significant profit in interest. I can't understand why anyone will claim that this must have been the plan from the issuer all along; it would be the worst theft in the history of mankind when the thief ends up with less money than before they committed their theft.
.b