iiinova — where your digital self keeps evolving

Bryan Meyer @doingmath.bsky.social · 02/04/2026

I’ve got a couple lessons you might find useful as next steps w this group of students in the “Global Forecast” unit. See “Are They Related?” and “The Emissions Equation.” Maybe also “Degrees of Impact” docs.google.com/document/d/1...

docs.google.com

5 - Global Forecast

GLOBAL FORECAST How can we use mathematics to make predictions about climate change? ESSENTIAL CONCEPTS and PRACTICES 5.1 Building Rules for Exponential Functions Can I explain why exponential ...

011

Reposted by Bryan Meyer

Paris Marx @parismarx.com · 15/02/2026

OpenAI ”acknowledged in its own research that LLMs will always produce hallucinations due to fundamental mathematical constraints that cannot be solved through better engineering, marking a significant admission from one of the AI industry’s leading companies.” You can’t trust chatbots.

computerworld.com

OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws

In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limi...

2221261016

Bryan Meyer @doingmath.bsky.social · 31/01/2026

Hi @comapmath.bsky.social ! I’m looking for print Teacher’s Editions of your high school ‘Modeling our World’ series, but don’t see them available in your store (or anywhere). Know if/where I can find them?

000

Reposted by Bryan Meyer

Jonathan Hall @studymaths.bsky.social · 18/01/2026

Specialise, Conjecture and Generalise. A three-step plan to encourage and develop mathematical thinking. I'm going to use this thread to share a few of my favourite tasks that lend itself nicely to this idea. It would be amazing if everyone added their own in the replies.

A poster demonstrating how you can use a three-step plan of specialising, conjecturing and generalising to solve mathematical problems. The poster is split into four areas:

1) The classic problem "How many squares are on a chessboard?"
2) Solving the problem with much smaller boards.
3) Arriving at a conjecture involving square numbers.
4) Generalising the problem ofr all board sizes.

53411

Reposted by Bryan Meyer

Ilana Horn @ilanahorn.bsky.social · 09/01/2026

In this editorial, I explore what it means to teach (mathematics) in a time of polycrisis. I use Anna Tsing's argument from "The Mushroom at the End of the World" to unearth the capitalist logics that underlie much of our educational enterprise to reimagine new possibilities. #iteachmath #EduSky

3177

Bryan Meyer @doingmath.bsky.social · 19/12/2025

I figured you were busy wrapping up classes before break. I’m not doing much over break, so we can connect then if that works for you.

010

Bryan Meyer @doingmath.bsky.social · 19/12/2025

…Because the LSRL always goes through the point (mean(x),mean(y)), we can do this same thing for relationships that don’t start at (0,0). Was wishing I had your applet built for my data set! Thanks for the ideas you always share here!

110

Bryan Meyer @doingmath.bsky.social · 19/12/2025

Crazy that you did this today! I was thinking about this EXACT applet that I’d seen from you when I was building a lesson on bivariate data…

110

Reposted by Bryan Meyer

The Maths Bazaar @themathsbazaar.bsky.social · 18/12/2025

Unexpected email from Jonny Griffiths telling me there's a new 2026 update of his RISPS book, that's great investigations for impressionable A Level students. Download it now from www.risps.co.uk with links to Geogebra and Excel files @jonny-griffiths.bsky.social

13011

Bryan Meyer @doingmath.bsky.social · 18/12/2025

I guess 'relationship' probably means something like: "the degree to which changes in one variable predict the changes in the other"?

000

Bryan Meyer @doingmath.bsky.social · 18/12/2025

But...wait! If you give the data trend even the smallest 'slope' the r value all of a sudden jumps to perfect. So, seems MAE is actually a very good predictor except in this 'anomaly' case where the data trend is horizontal?! Someone help! cc: @triangleman.bsky.social @averypickford.bsky.social

200

Bryan Meyer @doingmath.bsky.social · 18/12/2025

Ah! This extreme case is interesting. Here is a situation where the MAE is basically zero (VERY little 'error of the model') but the r value indicates 'no relationship.' So, measuring the 'error of the model' is not the same as measuring the 'strength of the relationship'

100

Bryan Meyer @doingmath.bsky.social · 18/12/2025

The internet says this: "In summary, while MAE and MSE relate to how spread out points are from a specific line of best fit not the overall variability of the raw bivariate data." But I don't buy that. Both MAD and StdDev are also in reference to central 'structure' of prediction (the mean!).

100

Bryan Meyer @doingmath.bsky.social · 18/12/2025

Hey #statistics friends, I need help with a question: With single variable data, we typically use MAD or StdDev as measures of variability. Why don't we just use MAE ('mean absolute error') or MSE ('mean squared error') as measures of variability for bivariate data?

100

Reposted by Bryan Meyer

Ralph Pantozzi @mathillustrated.bsky.social · 17/12/2025

“When do our students really understand the algorithms we teach? When they can use them effectively, or when they understand how they work, and can reconstruct them if they forget crucial steps” dylanwiliam137385.substack.com/p/onions #iTeachMath

A square table of 25 numbers

5
11
3
14
7
13
19
11
22
15
16
22
14
25
18
20
15
7
18
11
7
13
16
9

031

Bryan Meyer @doingmath.bsky.social · 11/12/2025

Thanks for this! Excited to learn. I have a couple specific questions (if you have time): - what is the value of being able to model sound waves w functions? - is there a meaningful use of solving trig equations in this context?

100

Reposted by Bryan Meyer

Dave Kung @dtkung.bsky.social · 01/12/2025

Is the point of mathematical research to know *what* is true or to know *why* it's true? Emily Riehl lays out a strong argument - and what it has to do with AI - along great advice for the field in @science.org www.science.org/doi/10.1126/...

science.org

Mathematics is hard for mathematicians to understand too

At a recent conference on mathematics in the age of automated proofs, mathematician and Fields Medalist Akshay Venkatesh presented “How do we talk to our students about AI?'' He quoted an email he'd r...

12510

Reposted by Bryan Meyer

Charles Logan @charleswlogan.bsky.social · 10/12/2025

"Big Tech and data center developers appear to be siting data centers in vulnerable communities...These locations show a troubling trend - they primarily impact working class and Black and Latine communities."

kaporfoundation.org

The Unequal Burden of Data Centers - Kapor Foundation

An examination of the Environmental and Public Health Impacts on Communities in California

512975

Bryan Meyer @doingmath.bsky.social · 10/12/2025

A couple things I’ve learned: - know board/district policy and use it to force them to do what it says - community needs to be educated on the issues and mobilized to show up - ORGANIZE; find ways to coordinate communications, efforts, and actions so that everyone interested in working together

010

Bryan Meyer @doingmath.bsky.social · 10/12/2025

Anyone have ideas about connections between sound/music and trig functions? cc: @averypickford.bsky.social @maria-naturalmath.bsky.social #iteachmath

300

Reposted by Bryan Meyer

Jess Calarco @jessicacalarco.com · 19/11/2025

Humans are bad at big numbers. Which makes it easier for billionaires (and trillionaires) to get away with hoarding wealth. Because most people just can't comprehend how much money that is. So, I appreciate this WaPo effort to help people visualize what ridiculous amounts of wealth really mean.

Elon Musk is set to make more than every U.S. elementary teacher combined
See how your profession stacks up against Musk’s pay.

By Alyssa Fowers and Leslie Shapiro

Every elementary school teacher
On average, Musk will make $3 billion more per year than the 1.4 million elementary school teachers in the U.S. combined.

Each figure represents 1,500 elementary school teachers. Together, they made $97 billion last year

315058

Reposted by Bryan Meyer

NCSM: Leadership in Mathematics Education @mathedleaders.bsky.social · 06/12/2025

Math education is complex. It can’t be reduced to single studies or one-size-fits-all claims. NCSM’s new position paper offers a balanced, equity-centered, research-informed alternative—one that honors the real nuance of teaching + learning. Read more → mathedleadership.org/position-papers

11210

Bryan Meyer @doingmath.bsky.social · 03/12/2025

Hmmm. Were people purchasing the curriculum because it had the Desmos name on it? If so, that feels troublesome.

110

Bryan Meyer @doingmath.bsky.social · 23/11/2025

Will you be posting your curriculum for the course at some point? (Or is it already there and I missed it?)

100

Bryan Meyer @doingmath.bsky.social · 23/11/2025

Could you see a way to do that with high school juniors or seniors?

000

Bryan Meyer @doingmath.bsky.social · 23/11/2025

This idea is the one I really want to focus on, but in a more societal context. Instagram recommending loop that leads to ideological bubbles, for example.

200

Bryan Meyer @doingmath.bsky.social · 22/11/2025

Any time-consuming procedure that is irrelevant to be fluent with in the age of the computer (factoring, solving complicated equations by hand, polynomial division, etc etc). So, 50% of the high school curriculum?

000

Reposted by Bryan Meyer

Robert Reich @rbreich.bsky.social · 21/11/2025

Thanks to his new Tesla pay package, Elon Musk could stand to make $3 billion more per year than all 1.4 million elementary school teachers in the U.S. combined. Still wondering if inequality is out of control?

9630871027

Bryan Meyer @doingmath.bsky.social · 21/11/2025

What I'd really like to do is center a unit on a context that allows for ethical considerations of the mathematical decisions made (recommender systems optimizing for time on app, for example) and that provides a good grounding for exploring the relevant central mathematical ideas (and practices).

000

Bryan Meyer @doingmath.bsky.social · 21/11/2025

I agree with what you've said. And both the fruit classifier and number classifier are fun ideas! Do you have these already built out into full curriculum units? (The slide examples are nice, but not enough for teachers to implement)

200

Bryan Meyer @doingmath.bsky.social · 21/11/2025

Thanks Tom! I want to build a high school unit for 11th/12th grade that uses basic matrix/linear algebra (context of recommender systems, ML, or genAI). Do you already have this in your course?

100

Bryan Meyer @doingmath.bsky.social · 21/11/2025

I want to learn more about how (high school) math is used in building AI. Any recommendations? @jedediyah.com @tombutton.bsky.social @dandersod.recursiveprocess.com

100

Bryan Meyer @doingmath.bsky.social · 12/11/2025

Conceptual understanding is not just knowing a principal, it must involve being able to explain why something is true. (See below from CCSS) I don’t see that present in any of the classroom examples given, really

100

Bryan Meyer @doingmath.bsky.social · 12/11/2025

I don’t see any advancement of teaching for understanding in the post, personally. Perhaps the opposite, actually.

210

Bryan Meyer @doingmath.bsky.social · 12/11/2025

What did you like about it?

100

Bryan Meyer @doingmath.bsky.social · 11/11/2025

The thing I don’t understand is why your post seems (to my reading, at least) dismiss conceptual understanding or aiming for kids to be able to articulate why something makes sense/is true. Am I off there? How are you thinking about that?

120

Bryan Meyer @doingmath.bsky.social · 11/11/2025

(b) Correct answers do not always equate to understanding. We have to ask questions to see how kids are thinking about the underlying ideas

120

Bryan Meyer @doingmath.bsky.social · 11/11/2025

In general, though, your closing feels pretty spot on to me… (a) Conceptual metaphors (the balance scale, ex) are useful IF we know what general principle/understanding we are hoping to develop (and gimmicky and disconnected for kids otherwise). And…

110

Bryan Meyer @doingmath.bsky.social · 11/11/2025

(3) “combine like terms” This is about equivalence, at the core. But (many) kids don’t know that bc all we do is teach symbolic manipulation. “What is the value of each expression when x = 1? When x = 5? Will any of these expressions ALWAYS have the same value? How do you know for sure?”

110

Bryan Meyer @doingmath.bsky.social · 11/11/2025

(2) f(x) increasing when f’(x) >0 Similarly, would want kids to understand WHY this makes sense (is true) not only THAT it is true. So, again a line of questioning like: “pick two points on f’(x). What do the coordinates mean? Based on that, when is f(x) increasing? Why?”

110

Bryan Meyer @doingmath.bsky.social · 11/11/2025

(1) a x b = b x a So I would want kids to have an explanation for why this must be true. I can imagine several “age appropriate” justifications, but a teacher (or curriculum) would need to ask something like: “is 5 x 6 the same as 6 x 5? 3 x 7 and 7 x 3? Will this always be true? Why (not)?

110

Bryan Meyer @doingmath.bsky.social · 11/11/2025

Hey, Michael! The “why” feels practically inseparable from what I consider to be conceptual understanding. I personally find the provided definition to be very clear and practical. Here is how I see it as actionable, fwiw, using the examples in your post…

110

Bryan Meyer @doingmath.bsky.social · 11/11/2025

This is from the Introduction of the CCSS. What is your reaction to this definition?

100

Bryan Meyer @doingmath.bsky.social · 04/11/2025

Can I use this in a (publicly available) trig unit I’m building if I cite you as developer?

110

Bryan Meyer @doingmath.bsky.social · 04/11/2025

Behind paywall? Do you have PDF?

100