The blog post seems to be confusing being able to release software quickly with releasing quality software faster. Reading between the lines one might infer that Evernote has a new CI pipeline - that's nice, it means they can make releases faster - it ALSO means they can release mistakes faster. Using customers to do what should be done by automated testing and a QA test team is a recipe for disaster.
My bugs for the the windows client would include
crashes a lot or refuses to start because the previous instance is a hidden zombie
displaying blank pages for notes that i know had content the previous day
The blog post says that the majority of customers want new features, but I doubt that a majority want bedrock core features sacrificed for those new features. I totally understand a desire to convert free users to power users by adding new features. I'd argue that only works if you make those new features premium good enough that people will pay a premium for them. As an existing premium customer I've not yet found any new features that I'd consider premium and I cannot find bedrock features such as auto importing from chosen file system folders ( desktop or 3rd party cloud drives )
If you really do want new features maybe let your paying users vote for them - engage with your customers to find out weekly what they want you to focus on. You may find the new features they want are actually old features you removed. You could even build such voting into the new app if you can find a way to stop it crashing.