How can laser diffusion be reduced? So, a laser works by bouncing photons back and forth between two mirrors until they straighten each other out and exit a small hole, like this:

The problem is that no matter how much the photons straighten out and how small the hole is, it will never be perfect, and there will at least some diffusion.
My question is this: Once the laser beam has exited the device, how can the diffusion be reduced?
I haven't found anything online about how this is done, but I assume that it can be done since LIGO exists and I'm pretty sure they wouldn't just make the most concentrated laser they can and hope for the best.
Ideally, the process wouldn't cause the laser to be focused to a point, as this would still create diffusion, even if there is less of it; but beggars can't be choosers.
 A: There is a fundamental limit to how collimated (or "straight") a beam of light can be.  The limit is in fact due to diffusion.  There is a sort of uncertainty principle at work.  The more you constrain the size of the beam, thereby knowing it's location,  the less you know about the transverse (perpendicular to the beam) momentum.  That is, the narrower the beam, the more the light is allowed to go off at an angle.  The size of the angle is inversely proportional to the size of the beam: $$\theta = \frac{\lambda}{d}$$ and the size of the beam a distance $L$ from the source is $$D=\theta L=\frac{\lambda}{d}L$$
To reduce the diffusion and the spot size some distance away, you need a larger beam.
Example:  a HeNe laser with a beam size of 5 mm and wavelength 633 nm is aimed at the moon, the size of the beam on the moon would be 49,000 m in diameter.
Example: now lets buy a huge beam expander and expand the beam to be 1 m in diameter.  Now the spot on the Moon is 243 m across.
The beam can never be perfectly collimated.  There's always some beam spread.
A: There is no straightening! Lasers rely on the induced emission of photons, which means that every new photon is identical to the one that caused its emission. That is they are identical in their frequency, direction, mode structure, polarization, etc. 
There is indeed homogeneity due to various imperfections: the low quality of the resonator, Doppler effect in gas lasers, different level widths in atoms due to perturbations of crystal lattice, etc. This homogeneity varies depending on the type of the laser: lasing semiconductor diodes are among the worse, gas lasers (like Helium-Neon) are among the best, $H_2$ maser serves as the frequency standard.
