Single Ended DAC vs Dual Differential XLR DAC

You are right...being sloppy. If all you are doing to convert an SE signal to XLR is adding an "inverting amplifier" you will not get all the benefits of a device being "balanced" and depending on the quality of that section of the amp/premap/DAC, etc... you may get degradation in quality. This is where you can get the phenomenon of RCAs sounding better than XLRs in non-balanced equipment.

If the entire circuit is fully balanced, then it will be cleaner sounding but if the components that balanced signal is being sent to are not balanced then all that excellence is lost.

In practice, this shows up in DACs. Take the Rockna Wavedream SE vs. XLR. In a fully balanced system, the XLR DAC will outperform the SE. In an SE system, the XLR and SE will sound the same.

You will get a better s/n ratio and a cleaner, clearer signal with a fully balanced unit.  That being said, if you do not have a fully balanced system, it is unlikely that you will get the full benefit of this.  This is largely due to the fact that at some point, the balanced signal will be converted to a single ended signal when it hits the unbalanced component.  Common mode noise rejection is lost and the quality of that transformer will have a BIG impact on sound.  

That transformer is often why, single ended devices sound better when using RCAs rather than XLRs with a transformed signal.  It is also why balanced devices XLRs almost always sound better than the RCAs that typically accompany the units.

If your system is fully balanced, you should get a fully balanced DAC.  SE will never sound as good.  If your system is single ended, a fully balanced unit will sound as good as a single ended unit at best.  Pay a lot of attention to the quality of the output of the RCAs if you get a fully balanced DAC.  Brands like Bricasti and Weiss have virtually no difference in performance between XLRs and and RCAs in practical terms but those are more the exception than the rule.  Many others see signal degradation in that XLR to RCA transformer.  

The reason the circuit is duplicated, (4 chips vs. 2 chips, etc...) is driven by the need to produce the inverse for common mode noise rejection.  That extra chip does not increase resolution, etc...