The model for investing you describe still exists. However, we now call such high risk/high reward innovation-oriented investing venture capital. It could be that the distinction you are trying to make is one of scale -- small scale investors cannot provide enough VC to get most good ideas off the ground, and as such are restricted to investing in larger, lower risk enterprises. On the other hand, it seems to me that the main change over the past fifty years on this front is the distribution of potential capital, in that many more people now invest in the stock market on a per capita basis. From that observation, I would contend that the narrowing of risk profiles is a natural consequence.

Fortunately, the actual return induced by this behavior reflects the reduced risk. See, e.g., bear.warrington.ufl.edu/ritter/PBFJ2005.pdf.

The hemoglobin in your red blood cells is reasonably paramagnetic; under the application of a large magnetic field it will produce a magnetic dipole. I suspect that the effect they are describing arises when two red blood cells get near each other. Then, the magnetic field from the induced dipole in the hemoglobin gets them to line up, much like what happens with pairs of refrigerator magnets when you bring them close. This grows into a longer and longer chain, until brownian motion overcomes the weak binding induced. The resulting chains of hemoglobin flow past each other more easily than individual particles, so long as they maintain their narrow aspect along the flow direction. The benefit claimed in the article thus pertains primarily to flow along the magnetic field's axis, where the external field keeps them oriented along its axis.

It is unclear what the metabolic effects of such chains are in practical settings--for example, how well oxygen exchange will occur with much of the cell membrane locked up against adjacent cells. Also, perpendicular flow may have a lower or higher viscosity as the unmagnetized sample (though the article is not available for reading yet, so I can only infer that it is still a bit lower due to the statements in the news release-ish article that the effect persists for some time after the magnet is turned off).

The key idea is that of inflation: general relativity allows for the distance between points to increase faster than the speed of light. Alan Guth's theory for inflation proposes that this in fact occured in the early universe, and the theory is now backed up by observations of fluctuations in the microwave background radiation (among others), where microscopic fluctuations were "frozen in" due to the rapid expansion. The consequence of this inflation is that much of the current universe is not within our 14 Gyr lightcone.

As a side note, the big hub-bub about dark energy is that it appears (based on current observations) that our universe may be entering a second inflationary period. Fortunately, the timescale for this is on the order 100 Gyr, so it will be unlikely to effect our lives directly.

Actually, quantum cryptography can work with non-secured intermediate repeater stations. In essence, rather than attempting to send the random bits directly, one attempts to build an entangled pair of quantum bits, one at each end of the repeater chain. This is trying to build a specific state, which can be verified before use. The random key is generated using the non-classical correlations of the entangled pair (for more info, one can google "Ekert quantum repeater").

As you might expect, the protocol for this case is very different from that used in classical repeaters: one cannot measure the signal and amplify it, because doing so would negate the security you are attempting to establish. Instead, a quantum repeater focuses on the specific task of creating an entangled state at the either end, which can then be used to generate cryptographically secure random bits. Because one can verify the final state at the ends of the system before it is used, and independently of malicious users in the middle, a quantum repeater is no more susceptible to an intercept attack than a short-distance single fiber approach. Again, both are susceptible to a complete man-in-the-middle replacement / redirect.