An observation I've made is that DaVinci Resolve seems to be a more technically capable still image editing tool than Adobe Lightroom, despite being designed for video colour grading instead of photography.
For example, Lightroom does not allow custom colour spaces for the output, they're restricted to a fixed list of baked-in ICC profiles. The only option is to export in the large ProPhoto space and then externally convert to the target space as a second step.
Lightroom has no support for true HDR, the type that requires HDR10 or Dolby Vision displays. If you want to edit a 14-bit RAW photo for, say, the background image in a HDR computer game or a YouTube video, Lightroom can't help you.
Even in SDR mode, the colour grading in Lightroom is much more primitive than Resolve. There's no vector scope, for example, something that would be trivial to implement for them.
Adobe talks a lot about their "AI" features, but they're very basic. They've left a lot of low hanging fruit on the tree. For example, it would be great feature that when flicking through a bunch of shots of a face while zoomed in, the eyes should be automatically aligned between each shot to make it easy to compare shots. Similarly, they could train an AI to recognise in-focus and out-of-focus faces, and automatically "score" a shot based on things like: "Are both eyes in focus, or just one?"
These days, this stuff is trivial to implement in software, yet Adobe is still firmly rooted in the 1900s era of "It's just like your darkroom! All the terminology of the black and white film era is unchanged! You can dodge! You can burn! Just... with a computer!"
Reminds me of AutoCAD 20+ years ago: it was a digital blueprint software. More computer-native CAD with proper 3D modelling capabilities such as SolidWorks ate their lunch...