Metrics can still be gathered without annoying the user.
If the goal is automatic exercise detection, it can log when this happens automatically without notifying the user. It can also show on the screen without the buzz. The user doesn’t need to interact for the action to be successful and recorded.
If the goal is more user interactions with the device, this detection needs to work both way, detecting when exercise as stopped. At that point it can notify with how many calories were burned and ask if they are going to keep going, so it gets a user interaction without being totally pointless. As long as it’s not bugging someone between ever set.
Of course I’d argue that “number of user interactions” is a terrible metric that almost guarantees enshitification. If the device is still being worn, it’s being used. Who cares how much the user fiddles with it?
Is the C++ language enshittified? Can you use the "numbers must go up" argument to explain it? Of course you can say the committee has to add new features to justify its existence. But that just seems like a superficial explanation. Maybe overall the newer editions of the language seem like a good idea. But complexity explodes until everyone is just using their own unique subset of the language. Anyway I'd like to read a take on enshittification and C++, since C++ is a long-lived "project" and the standards committee and process is more open than FitBit, with an more interesting mix of stakeholders (with more of a principal/agent problem?).
Metrics can still be gathered without annoying the user.
If the goal is automatic exercise detection, it can log when this happens automatically without notifying the user. It can also show on the screen without the buzz. The user doesn’t need to interact for the action to be successful and recorded.
If the goal is more user interactions with the device, this detection needs to work both way, detecting when exercise as stopped. At that point it can notify with how many calories were burned and ask if they are going to keep going, so it gets a user interaction without being totally pointless. As long as it’s not bugging someone between ever set.
Of course I’d argue that “number of user interactions” is a terrible metric that almost guarantees enshitification. If the device is still being worn, it’s being used. Who cares how much the user fiddles with it?
Is the C++ language enshittified? Can you use the "numbers must go up" argument to explain it? Of course you can say the committee has to add new features to justify its existence. But that just seems like a superficial explanation. Maybe overall the newer editions of the language seem like a good idea. But complexity explodes until everyone is just using their own unique subset of the language. Anyway I'd like to read a take on enshittification and C++, since C++ is a long-lived "project" and the standards committee and process is more open than FitBit, with an more interesting mix of stakeholders (with more of a principal/agent problem?).