Well... the topics that are adressed there are sometimes bordering on stupidity (because crucial info is left out) and other parts are perfectly true (in the eyes of many).
For instance discussing bit depth as a 'single factor' isn't a wise thing to do.
Bit-depth and sample-rate are closely connected and should be viewed as a whole not as individual 'components'.
There is also a rather big difference between a format (44.1/16 is the best known format) and DAC 'resolution'
The real resolution is called ENOB and is generally lower than the 'specified' resolution due to practical issues as tolerances and noise.
The fact that a DAC is specified for 24 or 32 bits doesn't mean it can actually reach that resolution but it may very well accept that digital format.
For DAC's:
1 bit can sound great as long as the sampling rate is high enough and a special kind of filtering is used (noise shaping) after all DSD is one bit and there are many DAC's that work on 1 bit by using noise shaping techniques to convert 16/24 or 32 bit signals to a stream.
4 bit can sound equally great as long as the sampling rate is high enough bla bla bla as above...(MASH = 4 bit DAC but it could easily be a format but isn't)
8 bit .... same story still oversampling needed where the article raises the expectation this is not needed. 8 bit with a low sample rate simply isn't enough to properly describe the waveform.
16 bit ... again perfectly fine, oversampling is not really needed any more but still has advantages mainly for easier analog filtering.
24 bit already borders on what is technically feasable these days, oversampling is not needed at all but still has advantages mainly for easier analog filtering in a DAC.
32 bit utter nonsense but can be handy for calculations/processing and filtering e.t.c. but there will never be a commercial DAC that will actually really be 32 bit.
One may say... a but there are mnay 32 bit DAC chips around so it IS possible... nope it only means the DAC's accept that format and pretend to reach that resolution which in practice they can't. reason .... noise floor. It's something we cannot get around.
It cannot hurt to describe a signal more accurately but going beyond certain limits is ... well ... not really needed.
There are also different formats that are well known and simply 'describe' how samples and bits are 'packaged'.
Most used formats are 44/16, 48/16, 96/24 and lately 192/24 but there are other formats as well but not used as much.
In practice it is not the bit depth that is the real problem it is the quality of the recording itself.
Creating a DAC with >120dB S/N ratio is rather easy, creating a microphone + amplifier >80dB is already much harder for instance.
As almost every recording is made using microphones this is the limiting factor.
Having a S/N ratio of a DAC <150dB is therefore utter nonsense as we can only hear a few dB into a noise floor and most certainly not >20dB into a noise floor.
The noise is so relatively loud it drowns any signal. so S/N ratios above 120dB are pointless.
Here is more to read about it:
people.xiph.org/~xiphmont/demo/neil-young.html#toc_ggage for instance.
It is 'hydrogen' talk and not everone will agree but the research they do (and did) is generally well founded and because of this is not based on 'golden ears'.
I think it is in the best interest of MANY parties in the hifi and high-end world that the division between the 2 'camps' must always remain.
For that reason it is probably best to just enjoy music and don't worry too much about the equipment.