Well if tinybox is not tempting enough yet, the next version will have USB, then after that I'll be working on an anti-wook death ray
I'd love to weigh in here, but I only have a few minutes in my busy work day to post. There are HUGE philosophical design differences between the Nbox and Tinybox and while I agree with most all of the comments above, none of them give the whole picture.
To start, the nbox is based on a high current, very low noise op amps design first published in the late 80s. It's a fantastic "no-compromise" circuit, but obviously tears through 9v batteries. Part of that is just inefficiency in the 4x9v design; two of the 9vs (the "every other show" ones) see very small current drain because they are powering the FET in the active cables. Current draw is about 1-2mA per cable and two fresh 9v's would run all-day long. The other two (the "every show" 9vs) are powering the ICs in a +/- configuration, so the battery life is essentially the mA of a single 9v and the current draw is an order of magnitude higher, so they only last about 4 hours. It wouldn't be too hard to design a 'drop-in' power supply that could generate the necessary +/- power supplies (15v would be good for ICs and the FET in the active cable) and I imagine we'll see a solution for this in the next year. I'm working on it when I can, but power supplies are not my forte by any means whereas Jon could probably sketch it out in about 5 minutes.'
The Tinybox, on the other hand, is a modern low-current design (single-supply, jon?) and is plain-and-simple apples to oranges. Personally, every experience I have had with the Tinybox has been good and while it's not an ultra low-noise design, it's still extremely clean. If sonic comparisons are going to be made, I think it's really only fair to consider that Jon does make a high-current, low-noise Tinybox and is philosophically more in-line with the Nbox design.
A few other considerations I think are worthwhile (probably more so than the switch or input/output connectors on either box) would be the design of the cables themselves and the intended use case. The Nbox cables generally come with very thin mogami cable and are infinitely better for stealthing. Genuine KC or KCY cables come with much beefier cable and overall their construction is far more robust, but that makes stealthing riskier and cumbersome. I don't stealth at all anymore, so my preference would almost always be the KC or KCY cables.
I also think that customer service is a strong consideration. Although you will never find anyone on this board say anything negative about nick, maker of the nbox, and his customer service, there will always be a difference between working with a legitimate company with a permanent fixture in the marketplace versus an entrepreneur/hobbiest. When you buy a pair of Schoeps cables, you are essentially paying up-front for a lifetime of guaranteed repair service at-cost. Schoeps will nearly always service used equipment and as a legitimate business can proprely set and meet most expectations. On the flip side, Nick has a real life and might now always be around. It's just like what happened to Bill and Rick (the creator and former 'nbox' makers).. they decided other things were more important in life and as a result, were no longer available to service existing gear. Again, I can't stress enough how good Nick has been in providing customer service, but that doesn't mean it shouldn't be a consideration.
One thing to consider is that other than the thinner cable of the nbox cables, there is really no advantage in their circuit design or sonic signature, at least not from my perspective. They're incompatible for preamp/powering design reasons and if you really needed to go stealth, a route you could go would be to take a pair of KC or KCY cables and have the stock cable replaced with the mogami thin cable. You'd need someone competent to do this, but it's not hard to disassemble the collettes and change out the cable. Needs to be done cleanly w/ careful attention paid to removing flux from soldering (although there's really not any soldering being done to the high-impedance side of the FET), but otherwise is straight forward. The design of the schoeps colettes is very slick, making it an easy job.
Edit to Add: One thing to mention regarding the nbox design.... and this may shock some of you nbox fluffers, but the THD performance of the IC's in use is really only 'good enough' for 16-20 bit recording. Better ICs do exist for recording in the 24 bit world, but I'm not sure what more you could expect from a 20+ year old design.