"While I am pretty pessimistic about most applications of generative AI to education, I am pretty agnostic about general applications. I am not rooting for this technology to fail."
This is an important nuance to which I say "amen." There really are some powerful insights arising from LLMs, and some interesting use cases -- it's just that neither of those are true for using them to tutor kids.
In this lesson the kids really have to use their eyes to notice things. Other than dropping a sheep they have no idea what will happen UNTIL THEY DO SOMETHING. That is the key. Letting kids experiment until they figure things out. They see the expression on the right. Then they have to notice the numberline along the bottom but after that they just drop sheep in different locations( I think...as I haven't done it either)and see what happens. As a K-5 teacher and specialist I tried to make all my lessons work this way because YES! it is a important way to engage students in their learning. All kids can drop sheep and see what happens not just the 5%.
Lollll the google ad! It’s depressing to think of a world being encouraged to turn to AI for matters of the heart, all the more reason your great work is so valuable!
Always fun reading your 95% series :) And thanks for the coteach.ai shoutout! I'd be curious to hear your thoughts. It follows the principle of co-designing with educators (similar to the AI remediation with experienced math teachers work you covered before https://arxiv.org/pdf/2310.10648), expanding to math scaffolds!
I have really been enjoying your series on the 95%, so please don't see this as criticism when I say that in my classrooms, I don't see a 95-5 split. I see a 5-30-30-20-15 split. And I mention this only because it adds yet another element of complexity to the teaching life. Thanks for all these posts. I've been trying to formulate my thoughts for an article, and they've been very illuminating.
When I’ve used this lesson I, and probably most kids who are used to school math assumed our goal is to eat all the grass without dropping in the water. And once they’d done that they’re ready to move on to the next screen. I love how offering an alternative thing to do like drop the sheep straight into the water before telling kids to try different things gives a way to start and emphasizes creativity. I wanna name it something like “opening the sandbox” but it’s more about noticing the sandbox, or stirring the sand?
It also feels like getting kids to notice and interact with what’s “not” not just what is. Like not just where is the grass, but where is the water?
Reminds me of doing the function machines lesson and telling kids “see if you can break the machine” “anybody find something else that gives you an error message?”
This activity allows the learner to break away from the fear of being wrong and allows them to look at the behavior of the inequality they choose to enter.
"While I am pretty pessimistic about most applications of generative AI to education, I am pretty agnostic about general applications. I am not rooting for this technology to fail."
This is an important nuance to which I say "amen." There really are some powerful insights arising from LLMs, and some interesting use cases -- it's just that neither of those are true for using them to tutor kids.
In this lesson the kids really have to use their eyes to notice things. Other than dropping a sheep they have no idea what will happen UNTIL THEY DO SOMETHING. That is the key. Letting kids experiment until they figure things out. They see the expression on the right. Then they have to notice the numberline along the bottom but after that they just drop sheep in different locations( I think...as I haven't done it either)and see what happens. As a K-5 teacher and specialist I tried to make all my lessons work this way because YES! it is a important way to engage students in their learning. All kids can drop sheep and see what happens not just the 5%.
Lollll the google ad! It’s depressing to think of a world being encouraged to turn to AI for matters of the heart, all the more reason your great work is so valuable!
Always fun reading your 95% series :) And thanks for the coteach.ai shoutout! I'd be curious to hear your thoughts. It follows the principle of co-designing with educators (similar to the AI remediation with experienced math teachers work you covered before https://arxiv.org/pdf/2310.10648), expanding to math scaffolds!
I have really been enjoying your series on the 95%, so please don't see this as criticism when I say that in my classrooms, I don't see a 95-5 split. I see a 5-30-30-20-15 split. And I mention this only because it adds yet another element of complexity to the teaching life. Thanks for all these posts. I've been trying to formulate my thoughts for an article, and they've been very illuminating.
When I’ve used this lesson I, and probably most kids who are used to school math assumed our goal is to eat all the grass without dropping in the water. And once they’d done that they’re ready to move on to the next screen. I love how offering an alternative thing to do like drop the sheep straight into the water before telling kids to try different things gives a way to start and emphasizes creativity. I wanna name it something like “opening the sandbox” but it’s more about noticing the sandbox, or stirring the sand?
It also feels like getting kids to notice and interact with what’s “not” not just what is. Like not just where is the grass, but where is the water?
Reminds me of doing the function machines lesson and telling kids “see if you can break the machine” “anybody find something else that gives you an error message?”
It's a nice opportunity to emphasize the truth that "you can't break math."
This activity allows the learner to break away from the fear of being wrong and allows them to look at the behavior of the inequality they choose to enter.