As it stands now, with the information available to us, I think the main point of C3 is a complete rewrite (yet again) to accommodate a browser-based application, rather than one that is coded to work directly on an OS level.
I understand why the devs are doing this: re-routing multi-platform support to a browser (Chrome) that already exists for all major operating systems, including mobile, is one reason (less work for devs). Secondly, the devs can develop with web tech, rather than C++, or whatever was used to develop Construct 1 and 2. Thirdly, having the editor run in Chrome allows them to implement live editing - for example, particle systems are editable live. Fourthly, it allows them to run Construct 3 as a service. Fifth, new functionality can probably be implemented faster for the devs this way.
Point 2 is a personal decision, and has nothing to do with the users.
Points 1 and 3 are useful to have, but many other game editors already support these things (well, no mobile editing on Android or iOS, of course). Personally I do not see the point of mobile touch screen only editing on small screens except for simple examples/prototypes, and the "coolness" factor.
Point 3 is already catered for in most other game engines worth their salt. Take Godot or Unity, for example: particle systems can be edited live, as is the case with many other things. Of course, when the output is web-only (Construct) it makes sense to switch to a browser-based editing environment for live editing: they both share the same foundation, which makes it simpler and more efficient for the devs to support and implement.
From my point of view, the majority of the benefits of C3 are initially aimed towards the developers themselves, with a secondary concern for the users. I am sure the devs feel quite differently about this. Certainly the benefits will trickle down more and more towards the users, and it is more convenient for the devs to roll out new features.
One issue, however, is performance in a browser. I have worked with browser-based applications, and compared to their desktop equivalents they always lose out in performance. A browser is yet another shell between the OS and the application, and it does have an impact. (for example, compare Gravit with desktop based illustration tools - way slower!)
I am sure the devs will add more features as time goes by - they should be able to roll out new functionality faster now that both the editor and the output share the same foundation.
But I am out. The rental model is unacceptable to me, and after a bit more pondering, I have come to the conclusion that it is a bad idea to rely on a specific browser (Chrome) as a foundation to build a game editor environment. If it would work cross-browser, I'd be more lenient perhaps.
Other issues for me are the lack of native export, no animation timeline, and other things such as the awkward handling of functions, no built-in translation support, and inconvenient control/input handling. Versioning and large project management are problematic as well.
Now that I am starting a fairly large game project I had to rethink my approach, and Construct 2 would not have been my choice anyway. With how the new Construct 3 is planned, I am convinced I would never even have considered C3. I think it is more aimed at small projects and not really meant to be used for semi-large scope projects. My opinion and expecation - but I can only know for sure when it becomes available.
Having said all this, I still find Construct 2 to be quite helpful for prototyping ideas quickly, or try out a quick movement setup. Perhaps that ought to be the aim of C3: the best prototyper in town