If that were the case, you wouldn’t just remember things a little wrong, you’d try and recall your name and instead be remembering a field trip you took in 3rd grade.
You are implying that the process must be error-prone? I don't see how that follows.
AIs can be checked too. And judged, tweaked, etc. Obviously
Groups of moderators can be just as biased as individual moderators. More so even. Given the amplifying effects of echo chambers.
Yes I read it. I just don't think it significates.
Lol. Imagine being the last human. Living in a bunker. Facebooking with your "friends"
I figure that which is written and rewritten is addresses. To memories and some kind of ambient idea-objects.
You don't even want to see an experiment? But that's the pudding.
But human moderators aren't perfect either. And they are often biased.
Oh nice phrase. Synonymous with smugnorant.
Wisdom and ignorance look alike in that there is a dearth of uncertainty.
Well giving that power to a human isn't so great either, clearly.
We already use text filters as a moderation bot. So we're just looking at improving the bot.
Maybe we're just looking for a way for the bot to recognize more complex patterns.
Ok, so given that memory is error prone, value makes more sense than reference, because errors of reference would be more errory.
That makes assumptions about how the referred-to stuff is arranged. It assumes no organization. That memories about lunch would be kept right next to the Chemistry lessons.
So, to step away from that assumption, maybe memory-components are more organized. Gradients of meaning, say.
And maybe memory-components are less chunky. Instead of a memory of lunch, it's memory of sandwich, table and chewing, arranged appropriately in the referencing data.