If this was five years ago, I'd probably be arguing on your side

I'm a big fan of Peter Aczel (TAC), and I'm usually inclined to agree with him (especially about things like cables), but in this case I can't. When I first started at Emotiva (a bit over a year ago), I brought my then-favorite Benchmark DAC1 - which I would certainly have considered to be one of those "modern generation" of DACs whose specs are way above anything that's worth improving on - in to the office; several of us listened to it against the old XDA-1.... and, yes, we matched the levels quite carefully (and the XDA-1 is also very flat, very quiet, and has very low distortion). Not to belabor the point, but the difference was obvious, to anybody who walked into the room, in two seconds flat, with almost any program material. The XDA-1 sounded "airier", it sounded like it had more high end; things were actually in different places IN the sound stage, and the sound stage itself was narrower with the Benchmark. One person present described it as "sounding like you were in a different
ROOM". Since then I've collected a few more DACs, and I can say with absolute confidence that many of them sound quite different from each other (and many sound quite similar). For all I know, the Benchmark may be more accurate, or maybe not, but there's no question that there IS a major
difference between them.
There's certainly room to question which is better, and whether the differences are important or not (or worth extra cost), but they clearly do exist. [My personal opinion is that they have something to do with the digital filters, which affect transients in time, and so can be different in ways that don't show up on traditional tests.... but I don't think anybody has done enough testing to say whether that's what's really happening or not.]
Another thing (which certainly makes sense) is that the differences seem to depend on the source material and what you're listening on. A good friend and I were listening to our new XDA-2 today - flipping the sample rate converter on and off - and playing a reissue of an old Herbie Hancock album at 192k. [Now, I'm one of the first people to agree with Aczel that we
shouldn't be able to hear any difference between 96k and 192k - but there was. Interestingly, neither of us could hear any difference using my AKG 702 headphones, which I would say have a pretty good high end, but it was very clear on my little Airmotiv monitors with their air-motion tweeters.]
Now, I'd not at all sure precisely what the difference was, or even which was "more right", but we both agreed that the cymbals sounded better with the SRC off, but my buddy was equally certain that the sax sounded more real with it on.
Someone (I think Einstein) is credited with saying something like "Make everything as simple as you can - but no simpler". In this case, I don't think the discussion is over. If you get the chance, listen to the XDA-2 (or an XDA-1, or a Schiit Bifrost), and then listen to a DragonFly, on a system or a pair of headphones with good high end; I'll bet you hear the difference pretty quickly. If not, then buy the cheapest one..... just don't be foolish enough to make the decision without listening for yourself.
I guess everyone has an opinion.
The Audio Critic and all their audiophiles, engineers, and PhDs would say that DACs have matured so much now that they don't make any significant difference at all.
And most of the DACs I've seen have great specs like SNR of -115dB or better and THD of < 0.05% or better.