Perhaps the most ambitious physics experiments of our age are the attempts to detect gravitational waves. Right now the largest detector is LIGO—the the Laser Interferometer Gravitational-Wave Observatory. This consists of two facilities: one in Livingston, Louisiana, and one in Hanford, Washington. Each facility consists of laser beams bouncing back and forth along two 4-kilometer-long tubes arranged in an L shape. As a gravitational wave passes by, the tubes should alternately stretch and squash—very slightly, but hopefully enough to be detected via changing interference patterns in the laser beam.
LIGO is coming into operation in stages. The first stage, called LIGO I, is supposed to allow detection of gravitational waves made by binary neutron stars within 65 mega light years of us. These binaries emit lots of gravitational radiation, spiral into each other, and eventually merge. In the last few minutes of this process you've got two objects heavier than the sun whipping around each other about 100 times a second, faster and faster, and they should emit a "chirp" of gravitational waves increasing in amplitude and frequency until the final merger. It's these "chirps" that LIGO is optimized for detecting. Later, in LIGO II, they'll try to boost the sensitivity to allow detection of in-spiralling binary neutron stars within 1000 mega light years of us.
To give you an idea of what these distances are like: the radius of the Milky Way is about 50,000 light years. The distance to the Andromeda galaxy is about 2.3 mega light years. The radius of the "Local Group" consisting of three dozen nearby galaxies is about 6 mega light years. The distance to the "Virgo Cluster", the nearest large cluster of galaxies, is about 50 mega light years. The radius of the observable universe is roughly 10,000 mega light years. So, if everything works as planned, we'll be able to see quite far with gravitational waves.
However, binary neutron stars don't merge very often! The current best guess is that with LIGO I we will be able to see such an event somewhere between once every 3000 years and once every 3 years. I know, that's not a very precise estimate! Luckily, the volume of space we survey grows as the cube of the distance we can see out to, so LIGO II should see between 1 and 1000 events per year.
The really scary thing is how good LIGO needs to be to work as planned. Roughly speaking, LIGO I aims to detect gravitational waves that distort distances by about 1 part in 1021. Since the laser bounces back and forth between the mirrors about 50 times, the effective length of the detector is 200 kilometers. Multiply this by 10−21 and you get 2 x 10−16 meters. By comparison, the radius of a proton is 8 x 10−16 meters! So, we're talking about measuring distances to within a quarter of a proton radius! And that's just LIGO I. LIGO II aims to detect waves that distort distances by a mere 2 parts in 1023, so it needs to do 50 times better.
Actually all this is a bit misleading. The goal is not really to measure distances, but really vibrations with a given frequency. However, it will still be an amazing feat... if it works.
But, the coolest gravitational wave detector of all—if it gets funded and gets off the ground—will be LISA, the Laser Interferometric Space Antenna: http://lisa.nasa.gov/
The idea is to orbit 3 satellites in an equilateral triangle with sides 5 million kilometers long, and constantly measure the distance between them to an accuracy of a tenth of an angstrom (10−11 meters) using laser interferometry. The big distances would make it possible to detect gravitational waves with frequencies of 0.0001 to 0.1 hertz, much lower than the frequencies for which the ground-based detectors are optimized. The plan involves a really neat technical trick to keep the satellites from being pushed around by solar wind and the like: each satellite will have a free-falling metal cube floating inside it, and if the satellite gets pushed to one side relative to this mass, sensors will detect this and thrusters will push the satellite back on course.