An engineering perspective:
The "carbon" question seems to be a recurring one throughout the forum. Although all have provided some great insight, one aspect of a cf seat post has been completely left out...the directional properties of carbon "fiber". Because we are all talking about carbon fiber and not just carbon, we must remember that the way the fibers are laid in the mix is EXTREMELY important to understanding the manner in which the post will react to loads (compression, bending, and torsion).
Most of the posts have centered around either cf seat posts making little or no difference, but always adding something referred to as "bling". All bling aside, these are my observations:
Carbon fiber bars (handle bar, seat post, toptube, etc.) are usually considered to be constructed with a lattice pattern of fibers that run around the bar and along its length. Some might say this is obvious b/c you can see that pattern, but that would just uncover a misunderstanding about cf. When you look at a cf bar, you only see the final layer of fiber - the "money layer". This layer is extremely thin and designed to catch your eye. The rest is hidden. For proof of this you should try and get a shot of a cf bike being constructed. I think PEZ has a recent article about this... Or you can just cut open your favorite cf bike's seat tube... The bike looks ugly until finished. TREK even uses a filler of sorts (aka bondo) to make those smooth joints you see before applying the money layer - the one you pay to see. The ramifications of this lattice structure is that the bar will twist some and bend some, but doesn't like to do both at the same time. Compression, well that's another story.
Compression (straight through the axis). No matter how a seat post is constructed (material wise) you will not be able to notice a difference in its ability to absorb shock in a direction parallel to its axis - if you have a very steep seat tube, straight down. Al, steel and cf are all fairly incompressible for the loading that we will likely impart on a seat post. The only way to change this is to buy a post that has a suspension mechanism. It will compress...
Bending. This is where cf should make its money. We all agree that a cf frame can be more pliable than an Al frame (I hope). This is because the resin is more flexible, and the directional nature of the fiber strands provide manners in which the carbon bar can (is allowed to) bend. If this weren't true, none of us would ride carbon forks. The truth is that Al forks are too stiff, and steel forks are too spongy. By putting a properly constructed cf fork on a bike we can control the material to only flex for bumps in the road but not when we are out of the saddle sprinting (side to side). If a cf seat post has the construction I explained above, it will bend a little (think for and aft on the bike) to make the ride more plush or at least remove some of the road buzz. Two things can make a huge difference though, the amount of bending force that is applied and the true carbon fiber pattern.
1. Bending moment. If you have your saddle all the way forward or aft and a lot of seat post exposed, the effects can be large, but likely not much more than a steel or Al seat post. Again, assuming the cf pattern is as stated above, the long fibers won't want to stretch or compress to allow movement. However, if your seat is close to the frame or you have your seat fairly neutral the effect will likely be that the post is stiffer that a steel or Al one - again the long strands don't like to stretch or compress.
2. The pattern. Now, if a cf post's pattern is not as I stated above (circular and lengthwise) then all of this can change. Below the money layer the strands can go in any and all directions. This means that the bending load may not be obstructed as much by the unwillingness of the strands to stretch and compress. Think of a mix or resin that just has a bunch of short fibers thrown into it. They are not uniformly directional, and therefore provide little if any directional characteristics. Therefore the bars willingness to bend will depend mostly on the fiber density. Now, the designer of the seat post has to start acting like a fork designer and lay the strands appropriately to achieve a seat post that is flexible in the right directions. This is what high end cf seat post designers do. That is why high end cf seat posts will make a difference, just like high end cf frames and forks. Stiff where they should be and flexible where they should be. This prevents bobbing while also eliminating road buzz.
Torsion (twisting). Unless you are doing some sort of funky riding or your seat post is so low your knees are crossing the top tube when you pedal, this shouldn't be an issue. Twisting is allowed by cf b/c the circular fibers going around the bar act like washers and the long strand going along the bar simply bend just a little (think of lightly twisting a bunch of dry spaghetti strands in your hands). However, I included this b/c it is the most likely cause of failure (other than deep scarring) of a cf post. When loaded torsionally and in a bending manner a cf seat post can fail quickly due to the multiple stresses being imparted (now think of the same strands of spaghetti being twisted and bent - now we have a recipe for disaster). If you have ever heard of cf handle bars failing, it's the same story. On the drops - or on the bar ends for MTB - a rider is twisting and bending the cf rods as the same time (probably why big time sprinters don't use cf - they are just too powerful for it). I snapped a cf bar on my MTB in this manner and learned the lesson the hard way. It's not always a rapid process either; you may only break a few fiber strands a week. But when it fails, you’re on your face in a heartbeat.
So that's what I think. Cf is here to stay because it is light, can be formed into any shape, and can be made to allow and prevent very specific loads and loading. Therefore, a cf seat post can make a difference if it is the right one. The problem is three-fold for the buyer. Too many cheap posts, not enough info provided by the good manufactures, and the cost associated with making a really good carbon bar (for any use).
Hope this helps. If not in your decision at least in your understanding…