I'll save you a whole lot of reading: no.
I repeat: no.
Still don't believe me? No, the DSLR is not dead.
Okay, you just can't take no for an answer, can you?
Using trailing 12 month retail sales numbers, here in the US the total dollar value of mirrorless cameras sold to customers was about a third that of DSLR cameras. Globally, that percentage changes with region, for sure. Plus the volume versus value numbers are getting distorted more as companies like Sony—and soon Nikon—emphasize high-end mirrorless.
In Japan so far this calendar year, both volume and value of mirrorless shipped into the country exceeds DSLRs by a bit. The Americas are the opposite, while Europe is closer to the US trend and Asia closer to the Japanese trend.
Looked at over time, there's little doubt that there's been a slow and steady shift towards mirrorless. As I've noted many times, this was inevitable, as from both a cost and manufacturing standpoint, mirrorless has benefits to the camera makers that DSLRs don't. I predicted that we'd see mirrorless equal DSLR volume probably in 2020, and nothing has changed in my assessment. Maybe we get there a few months earlier, maybe a few months later. But that's equal volume, not death of DSLRs.
Moreover, we have installed base to consider. DSLRs have sold in the hundreds of millions to date worldwide. Mirrorless won't reach that until the 2030's at the current rate of sale, if even then (predicting out that far is fraught with problems, so your margin of error grows). Clearly there is a large contingent still shooting with DSLRs, adding lenses, occasionally upgrading bodies, and more. That isn't going to go away any time soon.
Then there's this: Sony has upgraded their A7 line three times now. Yet I still would strongly advocate that the Nikon D850 (DSLR) is a better all around camera than the Sony A7Rm3 (mirrorless). They're about the same price. Which would I have you buy? ;~) And I can make a similar case for the Nikon D7500 (DSLR) versus the Sony A6500 (mirrorless) or Olympus E-M1m2 (mirrorless). That doesn't seem like death to me.
Still, the camera companies have a strong vested interest now to migrate DSLR owners to mirrorless. First, if they market them right, they might get a (mirrorless) sale they wouldn't have gotten from a satisfied DSLR user who wasn't thinking of buying any new camera. Second, manufacturing is simpler and thus less costly, and can be more automated. There's fewer parts to keep track of and coming JIT (just in time). Fewer of those parts are mechanical—which requires its own less efficient manufacturing method—and more are semiconductor-type that can be built automated with far lower product quality differences that have to be tested for. And while doing all that, the camera companies can easily eliminate some product bulk and weight, which coincides with market demands. Win, win, win.
For you, the photographer? Less clear wins. There is something to be said for two mirrorless focus modes, Single Servo focus and Face/Eye detect. For more static subjects, using the image sensor allows the first to have a confirmation step that DSLRs don't have and the second to be done computationally on information that's not as developed on DSLRs (the Nikon D5 DSLR generation does face detect, but it's using a lower pixel count imager in the viewfinder chain).
I mentioned the bulk and weight, though that runs both ways. DSLRs could be made smaller and lighter than they are without giving up anything important (other than more engineering cost for the camera makers—seeing a pattern here?). But there's also a benefit to bulk and weight as you start trying to do certain types of photography. One of the biggest issues with the Sony cameras (RX100, A6xxx, A7, A9) is that they've probably gone too far with size and compromised too much in terms of ergonomics and your ability to hold the camera steady or follow subjects reliably.
For sporting events, I'd still prefer a DSLR, particularly the D5, which isn't small at all. For wildlife, I'd still prefer a DSLR, partly because that's where the lenses are. For landscapes, yep, still a DSLR, particularly now that I've discovered how good the 19mm PC-E is (and Canon shooters say the same for the 17mm or 24mm TS-E). (Before people start harping on "but the EVF and focus peaking, Thom!", let me shut that down: the D850's rear LCD in Live View is better for landscape work than the A7Rm3's EVF or LCD, in my opinion.)
For people events, yes, that's where mirrorless has entered my production. Even shooting people in a studio with a tethered camera I might defer to mirrorless now. For compactness while shooting travel assignments, yes, mirrorless with the right lens set does provide advantages over DSLRs, and again, that's a place where mirrorless has entered and stayed in my gear bag.
Still, I think it safe to say that it'll be a while yet before DSLRs "die." It very well may be that DSLRs don't die any sooner than mirrorless does, they'll just target different uses. What do I mean by that? Well, we're getting mighty close to what I call ubiquitous shooting. That's a camera that is always capturing a very fast, high pixel count video stream and then doing interesting things computationally and ergonomically to create output, e.g. stills, that you've defined.
Ergonomically? Yes. Pressing the shutter release on such a ubiquitous beast would mean "save this moment, but analyze the moments before and after and either save some of those too, or pick/process the best." Heck, if this is a moving subject on a static background (even if you're panning), you can take the static portion from multiple frames to remove noise. And that's part of what I mean by computationally. The list goes on and on, and there's some pretty amazing things I've seen only in labs that I can't talk about that go down this line.
Until then, I'm shooting with DSLRs and mirrorless, and neither are more alive or dead than the other.
