The real danger is that eventually the app goes away due to not being maintained. Hopefully they release the API spec before then, but my specific device has already been reverse engineered at least.
I'm reminded of the Wemo netcams my parents owned, which updated to remove a previous ability connect directly to the video stream in favor of pay-only cloud software, and the proceeded to cease support for that, leaving my parents with two completely useless bricks.
I hope we do see phone/OS/store ecosystems that focus on longterm stability of features and UI. Keep devices repairable by end users, supply parts for the whole of the phone's lifespan.
Fair phone is trying to do this for phone hardware, and Framework is doing similar for laptops. We still need OS and store ecosystems.
And whether a device has controls doesn't matter with regards to hot water contact. It has to be properly sealed either way and capacitive buttons work fine for this (not worse than the low-quality buttons they'd use otherwise). The main feature you get with Bluetooth is another point of failure thanks to connection problems, at least that's my average experience with BT.
Making an immersion circulator app-only is likely a BOM-reducing measure, which is fine (given that active time is probably at most a few minutes like twice a day). And you can get smaller form factors too with app-only control (e.g. Ember Mug).
I like it, but now leave it pre-set to 135°F lest I set it elsewhere and then lose the ability to adjust it, so I think they dropped the ball on the design.
This is one of the few (unrooted) smart devices I actually appreciate. You can easily configure it for a specific task (steak? chicken? brussels?), get push notifications when it's done and even turn off it's warm setting remotely if needed.
And given that most of the thing is immersed in boiling hot water, it makes sense to not put controls on the device itself.