[hparadiz]

My teachers were frequently wrong too and spoke with authority on subjects in hindsight they were frankly ill equipped to teach. Part of learning is understanding how to reason through these types of issues. It's a common problem solving problem in the work place just the same.

[somenameforme]

I think this is a false equivalence, by some extreme margin. Teachers may occasionally have some fact or whatever incorrect, but especially at lower levels it's going to be exceptionally rare, and often based on a logical foundation that is otherwise invalid for some reason, like whether 1 is a prime number or not.

By contrast LLMs constantly get things wrong and once they get something wrong will begin weaving that into everything create entirely fake realities of the sort that is more akin to a schizophrenic than somebody being mistaken on this fact or that.

[Oreb]

My own experience from elementary school was that teachers being wrong was shockingly common. There is no subject in which I would trust any of my elementary school teachers more than a state of the art LLM. This was in Norway. I’m guessing you were privileged to grow up in a place with better quality elementary education.

[beepbooptheory]

Yeah but the teacher says you're wrong even if you're right. The AI tells you you're right either way. The former can facilitate an important lesson, the latter doesn't ever give you the chance to.

[hparadiz]

I've been doing things like accounting where I upload receipts and have the LLM adjust a Google sheet with the money balances. The error rate over the past year has dropped from occasionally to never. That is because there's sub agents now running that check the work. If you have multiple LLMs running with a 94% success rate but you throw them into a group that requires a consensus suddenly the number basically hits 99%.

[beepbooptheory]

We simply need to run sub agents on the children's learning, then we will maximize pedagogic efficiency to 99%.