>DSLR and mirroless are hard to beat indoors in low light conditions.
That's exactly the area where smartphones have been killing DSLR and mirroless for years now.
That's because of internal DSP processing, combining multiple frames, machine learning AI, etc, but to the consumer it doesn't matter: they get a clearer, more stable picture than what comes out of the DSLR/mirroless and with way less effort.
Some DSLRs have function of combining of multiple frames together as well (I think Nikon D500 and latest Olympuses as examples).
This is often called HDR.
The issue with combining images together is that it works for static objects well, but if things move -- it does not. So low-noise digital sensors still seem to offer much better results.
And certainly, startup-time (or app selection time) + focusing speed, is simply unmatched by phones compared to DSLRs or mirrorless with phase focus detection
I do think that Denoising images with AI/ML will be common place even in open source Image processing tools like Rawtherapee.
So DLSRs having APS-C or full frame sensors with lower megapixel count will do well if images are post-processed (or in camera processed) wit these AI tools.
In fact, I was thinking that buying a used DSLR from 2012 circa for 150$ bucks -- will yield similar results as a 2K camera or a 1k smart phone.
Phones are easier to transport/carry. That's has been their reason to take over the lens+camera systems.
But I think camera makers can make photo gear fashionable again :-).
I am working on some ideas in that area :-)
>The issue with combining images together is that it works for static objects well, but if things move -- it does not. So low-noise digital sensors still seem to offer much better results
The problem is mirroless/dslrs have much bigger sensors with slower readout, and much worse DSP capability than say an iPhone. They also use it much more conservative that a mobile phone marker too (which just cares to get a nice looking image to the casual user, not for fidelity and ultimate control).
So, mobile phones for low light can still get better post-processing results for moving subjects compared to any mirrorless/dslr "HDR" mode, through quicker intermediate shots taken and combined, and more DSP resources to devote to the task.
(Samsung, Google, and Apple also have much more money than Sony and Canon to spend on state of the art AI/ML applied research and DSP developers).
That's exactly the area where smartphones have been killing DSLR and mirroless for years now.
That's because of internal DSP processing, combining multiple frames, machine learning AI, etc, but to the consumer it doesn't matter: they get a clearer, more stable picture than what comes out of the DSLR/mirroless and with way less effort.