Also: these folks love to claim what they are not trying to do, while supporting it just happening, So they're not trying to replace teachers, but as a commenter said, "what if you don't have a math teacher?" and something tells me that These folks wouldn't say "GET A MATH TEACHER!" and oh, build a fund to do that? Nope, they'd "help..."
I'm reading Robin Isserles' _The Costs of Completion: Student Success in Community College"https://www.press.jhu.edu/books/title/12672/costs-completion about how the edu-philanthropists are "fixing" college students failing math in college, by completely ignoring the fact that they're arriving with minimal skills and pushing college placement for everybody and being extremely creative with statistics inspiring articles like this about
the legal requirements to "maximize" completing college level courses in the first year. https://update.occrl.illinois.edu/winter24/community_colleges_dera/index.html . No, our faculty aren't "buying in" (yes, it's always transactional for them)... *because it doesn't work for the students and we don't like failing the students.*
I grew very concerned with the "math attitude" of administrators when I taught. About 80% of our incoming, fulltime students would not be able to begin a typical STEM major without 1.5-2 YEARS of math remediation (assuming Calc I first semester, 1st year). I strongly suspect that there is a significant source of conflict for both Illinois and my state, Pennsylvania. The high schools are not even coming close to preparing these students for college math, the student body numbers are dwindling and the pensions were grossly underfunded for years. Keep in mind many community college instructors pay in to the state pension-I suspect the goal is to ensure enrollment at community colleges, ensuring people are paying in to the pension. The community college gets away with enrolling an "engineer" that can barely add, the pension gets a payment and federal funds and grants of all kinds follow these students that are literally operating at an 8th grade level. If the federal government prevented college enrollment for anyone NOT operating at an 11th grade level, this would be quickly solved-and the students remediation can occur at an adult night 9-12 program, as that is the ONLY place that should be teaching Algebra. Our nation doesn't want to own how we are failing these high school students so it is all swept under the rug.
Human idiosyncracies are a hallmark of memorable teaching...The isolation of AI on a screen can't replace the strategic importance of human interaction in community.
Roughly speaking, the challenge is to create a product that is tolerant of low / no experience teachers, that scales to take advantage of the capacity of veteran teachers, and that plays a supportive role in turning one teacher into another.
This is a challenge that is not unique to edtech, of course. Lots of software faces similar challenges with user growth.
What if we needed three *different* products instead? yes the boundaries would be unclear, but software has a bit of a process for that via "standard/pro/Pro+" in reverse, with Pro+ for the neediest case.
Again I am not stating this is easy or even desirable, I am trying to understand the limits of the various thought processes, as I do routinely: (yes, scientific inquiry :-) for instance:
I can take on "below-average math teacher" because I was definitely that when I once tried to become a middle-school teacher. And because I sucked at classroom management, there were days when I just handed out the worksheets to get some respite. Is that what AI-powered EdTech aspires to be, computerized worksheets? Because I think that's definitely within their capabilities.
WHile they say they don't want to replace teachers, if there is no teacher, then when the product is used, it is being used instead of a teacher. If the teacher is incompetent, it is attempting to "replace" something that teacher isn't doing, rather than figure out how to get the teacher to do it.
They do not understand what is to be valued in teaching, so what they "replace" is not teaching but passing something off that looks like it.
>>The thing he misinterprets about us is that he thinks we’re trying to replace teachers, and he thinks we don’t value teachers. <<
Okay - someone needs to get in the time machine to Sal's interview with Bill Gates because, Jesus Fuck, they certainly implied that. Heck - I pulled that quote from the transcript. I used it in a "Teaching STEM in a Postpandemic World" PD presentation and talked about how a part of our attention needed to be on AI-proofing ourselves in the eyes of people in their Central offices who never met a computer that they didn't want teaching kids.
I mean, Sal says that teachers will move further and further into the background as "facilitators," he and Bill share some words that show they fundamentally don't understand students or teachers and then laugh the idea of computers replacing teachers off nervously, like they just noticed the mic was hot.
It was Gates' Unconfuse Me from August of 23. Sal said about the same things (teachers stepping back to "facilitators" while AI steps up and becomes teacher) while he was on the book and Khanmigo goodwill podcast tour. IIRC, his appearance on Adam Grant's podcast had me cursing out loud at the idiocy on a run in hte morning.
YMMV, but "teachers -----------------------------------> facilitators" with AI teaching has been the chorus to the Khanmigo song all along.
Please don't lump all edtech together. Open-ended, creative tools REQUIRE great teachers to introduce, guide, prompt and support learners throughout the project-building process. You can't do creative projects like some of these - https://www.wixie.com/blog?id=5807 - without the amazing work of an educator.
I checked out the link, and I guess what I do in my classroom with tech could be described as "blank screen projects." My students use the Desmos graphing calculator for graphing, sometimes we use Excel, sometimes we use google docs. And I like these tools because they're open-ended, you can do anything with them. But I don't think that's what the proponents of AI-powered EdTech are aiming for, is it?
Can I get some help interpreting the Experience vs Test Scores graph? As experience increases, the standard deviation of test scores increases? Is that desirable?
Also: these folks love to claim what they are not trying to do, while supporting it just happening, So they're not trying to replace teachers, but as a commenter said, "what if you don't have a math teacher?" and something tells me that These folks wouldn't say "GET A MATH TEACHER!" and oh, build a fund to do that? Nope, they'd "help..."
I'm reading Robin Isserles' _The Costs of Completion: Student Success in Community College"https://www.press.jhu.edu/books/title/12672/costs-completion about how the edu-philanthropists are "fixing" college students failing math in college, by completely ignoring the fact that they're arriving with minimal skills and pushing college placement for everybody and being extremely creative with statistics inspiring articles like this about
the legal requirements to "maximize" completing college level courses in the first year. https://update.occrl.illinois.edu/winter24/community_colleges_dera/index.html . No, our faculty aren't "buying in" (yes, it's always transactional for them)... *because it doesn't work for the students and we don't like failing the students.*
I grew very concerned with the "math attitude" of administrators when I taught. About 80% of our incoming, fulltime students would not be able to begin a typical STEM major without 1.5-2 YEARS of math remediation (assuming Calc I first semester, 1st year). I strongly suspect that there is a significant source of conflict for both Illinois and my state, Pennsylvania. The high schools are not even coming close to preparing these students for college math, the student body numbers are dwindling and the pensions were grossly underfunded for years. Keep in mind many community college instructors pay in to the state pension-I suspect the goal is to ensure enrollment at community colleges, ensuring people are paying in to the pension. The community college gets away with enrolling an "engineer" that can barely add, the pension gets a payment and federal funds and grants of all kinds follow these students that are literally operating at an 8th grade level. If the federal government prevented college enrollment for anyone NOT operating at an 11th grade level, this would be quickly solved-and the students remediation can occur at an adult night 9-12 program, as that is the ONLY place that should be teaching Algebra. Our nation doesn't want to own how we are failing these high school students so it is all swept under the rug.
this blog is like the last stand of humans.
Human idiosyncracies are a hallmark of memorable teaching...The isolation of AI on a screen can't replace the strategic importance of human interaction in community.
{looks at today's presentation of molecular geometry containing J. Law Hot Ones appearance along with the structures of capsaicin and casein...}
Can verify.
Thanks for your cogent thoughts as always.
On both sides though, I am missing a discussion of the use cases:
* No math teacher at all (Tech is better than nothing?) - keep in mind the increasing recruitment problems worldwide.
* below-average math teacher, and/or neophyte early per your curves (Tech might boost up ?); see https://ssrn.com/abstract=4573321
* above-average to excellent math teacher (Tech helps keep track? used cautiously).
I'd love your comments or posts on these situations.
Hi Charles, thanks for the questions. The question about below-average & no teacher came up in a recent newsletter and I'll link my response here.
https://danmeyer.substack.com/i/156902547/featured-comment
Roughly speaking, the challenge is to create a product that is tolerant of low / no experience teachers, that scales to take advantage of the capacity of veteran teachers, and that plays a supportive role in turning one teacher into another.
This is a challenge that is not unique to edtech, of course. Lots of software faces similar challenges with user growth.
Thanks for responding, Dan.
What if we needed three *different* products instead? yes the boundaries would be unclear, but software has a bit of a process for that via "standard/pro/Pro+" in reverse, with Pro+ for the neediest case.
Again I am not stating this is easy or even desirable, I am trying to understand the limits of the various thought processes, as I do routinely: (yes, scientific inquiry :-) for instance:
https://curriculumredesign.org/wp-content/uploads/Benchmark-design-criteria-for-mathematical-reasoning-in-LLMs.pdf
I can take on "below-average math teacher" because I was definitely that when I once tried to become a middle-school teacher. And because I sucked at classroom management, there were days when I just handed out the worksheets to get some respite. Is that what AI-powered EdTech aspires to be, computerized worksheets? Because I think that's definitely within their capabilities.
I am not partisan of the Khanmigo approach; I am just asking Dan questions to understand the use cases better, and their limits.
So they're only trying to replace *these* teachers, right?
I am looking for a cogent conversation, and I wasn't born yesterday.... ;-)
WHile they say they don't want to replace teachers, if there is no teacher, then when the product is used, it is being used instead of a teacher. If the teacher is incompetent, it is attempting to "replace" something that teacher isn't doing, rather than figure out how to get the teacher to do it.
They do not understand what is to be valued in teaching, so what they "replace" is not teaching but passing something off that looks like it.
>>The thing he misinterprets about us is that he thinks we’re trying to replace teachers, and he thinks we don’t value teachers. <<
Okay - someone needs to get in the time machine to Sal's interview with Bill Gates because, Jesus Fuck, they certainly implied that. Heck - I pulled that quote from the transcript. I used it in a "Teaching STEM in a Postpandemic World" PD presentation and talked about how a part of our attention needed to be on AI-proofing ourselves in the eyes of people in their Central offices who never met a computer that they didn't want teaching kids.
I mean, Sal says that teachers will move further and further into the background as "facilitators," he and Bill share some words that show they fundamentally don't understand students or teachers and then laugh the idea of computers replacing teachers off nervously, like they just noticed the mic was hot.
It was Gates' Unconfuse Me from August of 23. Sal said about the same things (teachers stepping back to "facilitators" while AI steps up and becomes teacher) while he was on the book and Khanmigo goodwill podcast tour. IIRC, his appearance on Adam Grant's podcast had me cursing out loud at the idiocy on a run in hte morning.
YMMV, but "teachers -----------------------------------> facilitators" with AI teaching has been the chorus to the Khanmigo song all along.
Please don't lump all edtech together. Open-ended, creative tools REQUIRE great teachers to introduce, guide, prompt and support learners throughout the project-building process. You can't do creative projects like some of these - https://www.wixie.com/blog?id=5807 - without the amazing work of an educator.
I checked out the link, and I guess what I do in my classroom with tech could be described as "blank screen projects." My students use the Desmos graphing calculator for graphing, sometimes we use Excel, sometimes we use google docs. And I like these tools because they're open-ended, you can do anything with them. But I don't think that's what the proponents of AI-powered EdTech are aiming for, is it?
Can I get some help interpreting the Experience vs Test Scores graph? As experience increases, the standard deviation of test scores increases? Is that desirable?
Not a data scientist! so take this interpretation with a grain of salt, but according to the paper: https://files.eric.ed.gov/fulltext/ED618564.pdf p16
"we have normalized the test scores to have a mean of 0 and standard deviation of 1"
So the positive SDs are effectively a comparison to a teacher with zero years of experience?
So as SteveB says, it's a measure of how many SDs above a zero experience teacher, is a teacher with X years of experience.
I think it's using SD as a yardstick to measure how far above [something] the teacher is. How many SD's above ____
But I don't think it can be SD's above the mean, because this isn't Lake Wobegon, where all the teachers are above average.
Truth!
and then this quote:" designed to flatten the human mind". Sigh.