To explain a bit more: If a merchant receives fraudulent credit card transactions, he/she could reasonably refer to those as "not real" because once the banks discover the fraud, they have the power to move money and effectively nullify those transactions in many cases.
In contrast, on the blockchain (and off the blockchain for most bitcoin exchanges), that kind of thing doesn't normally happen. Even in Mt Gox's case where they directly manipulated the price from the inside, the people who bought and sold on the exchange still experienced "real" price movement, and as long as they got their money/BTC out before Gox imploded (which is another matter), those transactions and price movements were every bit as real as any other. From the perspective of other exchanges at the time it didn't matter at all whether the driving force for the trend was fraud or completely legitimate behavior.
Finally, if MMM's "net effect is zero" as you say, then we wouldn't be seeing the price movement that we're seeing (and a price "boosted only by the float" is no different from a price boosted by anything else), or perhaps MMM just isn't the primary driving force in the current trend.
But the float is much smaller than the total volume of the fraud. If MMM is doing $10b of fraud a year, then it's not boosting Bitcoin's marketcap by $10b as one might naively guess ('$10b fraud drives Bitcoin price growth!' scream the headlines), it's boosting the price by much much less, by just the float (which could be more like $27m).
I'm not sure what you mean by "real" here. Even if demand is created from questionable/fraudulent sources, the transactions are just as real as any others. That's not to say it's sustainable or even driven by rational behavior (it never really is). Any trend, bubble or not, is just that - a trend. There may be numerous factors driving it, and it's almost never a good idea to make trading decisions based on the trend itself (the few exceptions of traders should be very aware of the limitations and risks involved in such a strategy).
> We'll see if the new fancy infrastructure that bitcoin got is rock-solid.
Are you referring to Bitcoin XT / BIP 101? As far as I'm aware[1], the support for this is still represented by a minority, so we're not really going to see any "new fancy infrastructure" tested, at least not for a while.
[1] http://xtnodes.com/