It is all very logical, and can have 3 different causes :
1. the setup being corrupted during transfer. Can you please confirm whether or not you see the popup "The patchdump was successfully sent" in UnO2ControlCenter? If you see this message, we can rule out an interface problem, if you don't see this message, the SysEx transfer fails and this causes the issue you see.
2. a firmware bug. In that case I will be able to reproduce the same behavior on my setup if I use the exact same setup, and then I can find out what is going wrong. The main question here is: did you use a different setup file in the second video, compared to the first video? If so, please send me the 2 different setup files so that I can reproduce the problem here. If the setup did not change, it proves that the issue is not with the firmware chip, but the cause is rather ;
3. a hardware issue, like the EEPROM chip (= setup storage) which no longer works reliably. With standard Behringer or UnO firmware, this can remain unnoticed because that firmware has a very rigid structure of storing the different PC/CC/Note messages for each preset on a fixed location, and the setup content cannot influence the controller behavior. UnO2 firmware works completely different, it reads and interprets the setup content, which can contain a wide range of commands. If one byte in such a command is corrupted, the whole system can become unstable as you see in your first video. In your second video, the system no longer is unstable (you can select presets, activate DirectBank, the switch LEDs behave as expected...) but the preset content is corrupted.
Of course I understand that this looks very much like an UnO2 firmware problem to you, so please email me your setup files so that I can either rule that out or solve it asap. I will stop shipping UnO2 chips until I have been able to investigate this.