Hey fujii, I've worked at a few towers too, and in the absence of pilot reports, we'd frequently use ceilometer measurements when visual observation didn't match what the dew point was telling us. Non-convective cloud can be hard to pick if you don't have any visual reference to compare it to.
Nirak lasers are light beams. Laser stands for "light amplification by stimulated emission of radiation". I don't think anyone was talking about the old cloud searchlights, which were out of date in the 1940s and I haven't seen one in thirty years.
I've always reconciled cloud reports AGL as being target audience - two people care about cloud base: pilots who wish to remain below it, and pilots who wish to make an approach through it.
Pilots wishing to remain below cloud need to know how high it is above terrain to determine if they can satisfy their 500/1000FT minimum height rules and remain in VMC - this is easier if expressed AGL. Pilots wishing to approach through cloud need to compare the base with their approach minima, which are admittedly expressed on most plates both AMSL and AGL, but are usually defined in legislation and ops manuals as figures AGL. Then you have the radioaltimeter, which is only in AGL.
Finally, at most meteorological sites around the country, the observers are not pilots, not controllers, and don't prepare their observation specifically for aviation. For these people, it's easier to determine height AGL and the end user - you - can do what they wish with the information.