avturchin comments on For alignment, we should simultaneously use multiple theories of cognition and value

[avturchin]

I meant that a situation is possible when two AIs use completely different alignment methods and also come to different results.