The more I read about your LLM policies, I realize many of you are just like my teachers who drilled us in obsolete skills like mental arithmetic, but failed to teach us relevant stuff, like basic probability theory, because they couldn't envision how technological change would transform the world.

Basic probability theory has been useful to know since there was any basic probability theory to teach, so this seems like an odd case to choose for making a point about technological change. The reason why it isn't reliably taught must lie elsewhere.

You don’t think mental arithmetic is important, to eg avoid being ripped off in shops, negotiating business deals, etc.?

Mental arithmetic isn't obsolete. It's essential for being able to judge the likely accuracy of data you're presented with

I find myself using mental arithmetic every day for estimating bills, calculating tips, and more. Perhaps the most important is being able to quickly estimate order of magnitude for project time and cost. I despair to see so many struggle with it.

That makes me wonder, do you have any thoughts on ours? Would be curious to hear your feedback! www.lw.uni-leipzig.de/fileadmin/Fa...

Additionally, in education, people rarely consider opportunity costs. Isn't most of the debate along the lines of, "Yes, we need to teach them X, Y, and Z now," yet without replacing content that has become less relevant or even unnecessary?

There is also a difference between training for Academia and for jobs outside Academia. It seems conceivable that traditional papers will die out, and producing research results will become much more systematic than just writing text. In contrast, this seems less likely outside academia, or rather

I don’t know. A lot of teaching ultimately revolves around how to distinguish the believable from the unbelievable, which in my experience is quite helpful with LLMs.

People should be able to do simple mental arithmetic by the age of 8. A basic skill! Probability theory was important in *any* age. Yes, it should be taught much more -- I had no access to formal instruction 'til university and was scandalized grad stats proceeded w/o prob. foundations.

Self contradictory. The concern is that LLMs will be used to avoid learning probability theory