[link=mailto:44.1@24]44.1@24bit[/link] all the time.  I do note many professional studios are using 96 and I may start using 96 for my next CD just because.  One of the reasons to use higher rates is for soft synth rendition, but many synths (and FX), upsample anyway these days.  Be aware many of theoretical benefits get lost in home studios (noise in the electrical lines, less than ideal rooms, etc).  But if you have plenty of CPU power and storage it probably won't hurt.

Doesn't really matter if you don't record for DVD/Bluray. If you are going to down pitch anything than it might be a good idea to use 96khZ or higher.
 
EDIT: 24bit of course is the way to go. :)

With well designed A/D conversion, 44.1kHz recording is perfectly adequate.  The problem is that many audio interfaces have mediocre A/D converters circuits that introduce ripple, phase shift and distortion at frequencies well below 20kHz.  In fact, any audible differences between 44.1kHz and 96kHz recordings can almost always be attributed to the quality of the converter design.
 
For example, a cheap converter design might have a 20kHz anti-aliasing filter that actually starts adding ripple and phase shift starting around 5kHz.  By running that same converter at 96kHz, the anti-aliasing filter can be moved up to around 40kHz, which moves up the start frequency of ripple and phase shift to about 10kHz.  With less of the musical spectrum affected, the 96kHz sample rate will sound better.  Not because 96kHz is inherently better, but because the higher sample rate lessens the negative effects of bad A/D conversion.

Some of us (including Ethan Winer, IIRC) think 16 bit is adequate. (But I use 24 anyway) In my listening tests, 48khz will sometimes give a smoother sound with plug-ins. Any higher than that and the law of diminishing returns applies. The big jump for me is between 44.1 and 48. (But I use 44.1 anyway. Ha!)

Thanks for the input. Thanks for not getting too technical about it as  well. I was wondering who was using higher rates and if they were actually hearing results that compelled them to continue. My interface (komplete audio6) uses cirrus logic converters, Rick I would be interested in your take on those. I most likely will try a higher rate for a smallish project (less than 20 tracks) and see if 1) my computer doesn't chug along too much and 2) if it improves the sound of synths and plugins. I'll let Ethan Winer ponder on the other stuff. I caught myself listening to some Bowie stuff yesterday (R I P) and had to remind myself of some of the limitations of the technology of the mid to late 70s. My take away was that good  performances,mixes and arrangements trump all.

I've yet to see any double-blind tests that prove conclusively people can hear the difference between 44.1 and 96 on playback, but I'm always open if such a thing exists. 
 
However, I've demonstrated in workshops at the New Music Seminar, AES/Mix Nashville, and at Gibson's New York showroom that recording at 96 kHz can make an audible, obvious difference under specific conditions that involve "in the box" plug-ins and virtual instruments. However in this case, with SONAR you can simply upsample when rendering, and gain the benefits of 96 kHz recording in 44.1 kHz projects. SONAR is the only program that does this, and I'm proud to say it was my idea 
 
I prefer 24 bits to 16 not because there's a problem with 16-bit technology, but because in the real world of power supplies, circuit board layouts, thermal noise, "glue" components around the converters. and the converter tolerances themselves, you never get 16 "real" bits. Some "16-bit" devices have an actual 14 or even 12 bits of resolution. However, a 24-bit system will virtually always give you at least 16 bits of real resolution, and usually more, like 18 or 20 bits.

Rick is correct about bad convertors, however most modern delta-sigma convertors internally upsample so the problems he describes are not as prevalent today.
If you have old converters that introduce phase issues at 5Khz, get another interface!
Your Cirrus convertors should be fine.

There is a reason some converters cost $500 and some cost $5000, but the difference between top of the line and average converters isn't as obvious as it once was.  Cirrus Logic has been in the converter business for a long time and certainly knows what they're doing, but the power supply, surrounding analog circuitry and quality of components also play a big part in converter quality.
 
If your interface sounds better at higher sample rates, then record at higher sample rates and do your final sample rate conversion in-the-box when mixing down, or let the mastering house do it.  If you can't tell much difference with higher sample rates, then your audience won't be able to tell either, (or won't care) and will like, or dislike, your song regardless.

As far as I know,
There are pros and cons.

I'm not a specialist or scientist but have been internated in technical things about sampling rate so I searched and read articles, forums, books.
I think Aliasing, EQ, IMD and SRC, are the important things.

For aliasing,
compressor, limiter, distortion, saturation and these kinds of process usually produce upper frequency information, and this could go above nyquist frequency, which will result aliasing noise. So, technically, with higher sampling rate, there will be less aliasing noise when you are mixing or mastering because nyquist frequency is set higher.

Also, when you tweak EQ near the nyquist frequency, the EQ curve doesn't be like you'd expect. It acts differently when it gets close to the nyquist frequency. (in a bad way i think)
So if you think about these, higher sampling rate like 96khz can be better.

Some plugins internally process at doubled of multipled sampling rate (oversampling) to avoid aliasing or that. The higher sampling rate is better in some case.

However, whenever you use non-linear processes like I mentioned above (compressor, distortion, etc.), they also produce IMD.
If the original material or recorded track has
a lot of information that we can't hear, lile above 20khz, these unnecessary information produces more IMD. So this is a downside. To avoid this, you can use a filter to cut these frequency before you put non-linear process.
So with higher sampling rate with unnecessary information, you might get even worse result because of IMD. I read this thing on gearslutz forum.

For SRC, I think you eventually release your songs at 44khz, so if you produce your music at other sampling rate, you need to convert the sampling rate with SRC. You can use SRC on your DAW but the quality of SRC is different on each program. And whenever you use SRC, the sound get worse and peak changes. (only just slightly and no one might notice though) anyway, usually SRC in your DAW is not ideal, so if you don't have decent SRC, the track converted to 44khz may sound worse than one with 96khz. IFAIK SRC on ableton live is great among DAW. SRC by Saracon or izotope are considered to be best I think. You can check the quality of SRC on SRC comparison website.

Reverb or other process may act differently on different sampling rate but I just don't know it.
Intersampling peak at higher sampling rate is beyond this topic I think, too.

So the end result will be really different not because you record it on higher sampling rate but you processd it on higher sampling rate.

I currently master at 96khz with ultrasonic filter to lessen IMD but for other thing like mixing or producing. I don't know which is better because higher sampling rate seems to require more CPU.


So, for producing, mixing mastering, there will be some difference in sound. But for recording, there might be subtle differences technically but these might not be significant.
If you mix at higher sampling rate you can record them at higher sampling rate to avoid unnecessary process.

Mixing down many 44.1 tracks to 44.1 will introduce evil.  Evil is bad.  Avoid evil by recording at 24bit/48khz.
 
Unless you have a REALLY REALLY SUPER-DUPER FAST !! system, forget going higher.