You have been here for 4 years now with over 10,000 posts read. I am sure you, like everyone else, has been following the forum chatter for the last 6 months.
Please provide one example of anything that wasn’t\isn’t complicated for Wyze to fix. Theoretically it shouldn’t be that hard, realistically it is Wyze. There is an old joke I remember about the difference between theoretically and realistically… Maybe a topic for the water cooler.
As it stands right now Wyze AI is completely unreliable. And, as a critical component of their core security model, it is unacceptable. I have more faith in the weatherman when he predicts a 50% chance for rain… It may rain… It may not.
As a byproduct of my 28% delivered notification rate issue:
I also logged the number of AI tagged events I received from my cams over a 1 week period. Less than 50% were correctly tagged. If the AI incorrectly identifies objects 50% of the time, isn’t it also logical to question if it also fails to tag objects at a very high rate? Instances of PD fails on the forums is well above any acceptable level. I would be very interested to learn the results if someone with the time and the resources were to run their long term SD Card continuous history thru a competitors AI engine. It is my prediction that Wyze would be embarrassed.
Unless Wyze is visited by the AI Fairy tonight and it leaves a solution under the pillow, I don’t predict any improvement soon. Wyze is too busy developing the IoT Toilet Paper Spindle.