So, I do a little bit of related research in this area: This is typically called a "filament", a "plasma channel", and sometimes described as a waveguide. It

So, I do a little bit of related research in this area: This is typically called a "filament", a "plasma channel", and sometimes described as a waveguide. It requires high powers (~10^13 W/cm^2, usually femtoseconds in duration) depending on gas. It essentially works because the index of refraction has two nonlinear terms: n(ω) = n_0 + n_2 I - n_plasma the n_2 I makes it so that higher intensities (which occur closer to the center of the filament for a typical gaussian laser beam), have a higher index of refraction, causing the field to focus, increasing the electric field intensity at the center. However, when the field gets high enough, it ionizes the material, causing a plasma to form, which has a negative contribution, defocusing the filament. These two contributions balance out and cause a "waveguide" where the light essentially travels along a narrow channel, allowing the focusing pulse to travel longer without as much spatial dispersion. You can't send another light beam along these channels, because after the beam has left, all that is there is a plasma channel, which defocuses and absorbs light instead. This has been around for a while. What is NEW here is that they modified their beam profile to have 4 or 8 or even an annular ring of these plasma channels. Accoustic waves propagate toward the center of this filament arrangement causing a change in the density (which changes the index of refraction, the index of refraction increases linearly with density) that makes the channels act like a waveguide for another light source. This source of the density change has a duration of microseconds. When the acoustic waves have dispersed, a thermal gradient has a similar effect, and this has a duration of milliseconds. They used this method to channel light from a spark source while mostly maintaining the spectra of the spark. Now, what this allows, or doesn't allow: LIDAR is a possible use, but whatever you point this source at is going to have to be able to take 10^13 W/cm^2... and not a lot of materials can take that without breaking down. While there is good focusing and guiding of the light, I think single photons are probably going to have a high probability of being absorbed (the mode of the photon will overlap into the plasma channels, which will likely absorb it), The nice thing is that signals could likely travel through one of these channels (though there likely will have to be some dispersion compensation at the receiving end).