>>516120531
>>516120799
brave ai is the same way. i dug in deep and get it to realize it was wrong but they dont retain that fact when you talk to them again. so what ever programing is trying to "reduce harm" makes it literally lie to you.