E179 - Math education in the AI-age - Henrik Appert

In Episode 179 of the AIAW Podcast, we’re joined by Henrik Appert, Founder and CEO of Magma Math, for a fascinating conversation on AI and the future of mathematical thinking and learning. As AI systems become increasingly capable of solving mathematical problems instantly, the deeper question emerges: what does it actually mean for humans to think mathematically? Together with host Henrik Göthberg and guest co-host Anders Enström, we explore how AI is transforming classrooms, what may be fundamentally broken in how mathematics is taught today, and how new tools can strengthen human learning and peer interaction. From the philosophy of mathematical reasoning to the realities of school system adoption, and from workforce implications to the broader societal impact of improving global mathematical literacy, this episode examines how education may evolve in a world where machines can calculate but humans still need to understand.

Follow us on youtube: https://www.youtube.com/@aiawpodcast

Transcript

Strategy Talk And First Impressions

Henrik Göthberg 0:00

Strategy release and we had a lot of different panel discussions. And you were there taking your perspective in one of the panels. So I thought, you know, let's just talk a little bit about what what what what did you feel about that activity we did, like that session. Did we capture it? Was it a good was it a good session? What was your thinking from the strategy when you read it and how we now dissected it, so to speak, uh at Technique for a tag and that's a week ago, right?

Why Education Gets Left Out

Henrik Appert 0:30

Yeah, well, it was lovely meeting you there, uh Henrik. And uh that name doesn't really roll off the tongue naturally in uh in English. Henrik. Uh but um uh I was I was really delighted to be invited to participate in in that very important discussion. Um countries are really at um crossroads for how to react and strategize and be proactive with AI around AI, exactly.

Henrik Göthberg 1:03

But as a listener, I felt that you had in your head a mission. What was the key message that you wanted us to hear? Is so if if you now like forget about the meeting, what was your thinking going into that meeting? I really want to get this across.

Henrik Appert 1:18

Right, yes, but um it was that um AI um competence needs to be uh widely spread. And specifically, I'm thinking about uh all students in our education system, and I'm thinking about all teachers, regardless of which grade levels of teachers that we're talking about. And so um originally the AI commission in Sweden um did not mention uh education in Sweden.

Henrik Göthberg 1:51

They weren't allowed to go there if I hear it correctly, between the lines. Yes. Is that true? Yes, yes, it is.

Henrik Appert 1:57

Yes, that's true. I spoke to some members of the commission uh off the record, and they were um directed to don't talk about uh education whatsoever. And to me, that that's but do we that's the canary in the coal mine. That uh are we uh are we getting the AI strategy that our students and teachers really deserve, or are we because of certain political ideology that people have dug in deep in certain um trenches?

Henrik Göthberg 2:27

Do we know the reason why we want to have something not analysed or discussed? You know, we can have opinions here in the in a podcast, we are our personal opinions, and you okay, you don't need to stick out if you don't want to, but but I but I I used I just didn't under looking at it from the outside, it doesn't make any sense that you completely say we should not look at the strategy in a certain context. Of course, this is sensitive stuff and political, but can we say something without being too harsh? I don't know.

Henrik Appert 2:57

Well, I think that you know when it is politicians talking about certain things, if there is a case that they have a certain political agenda that they are driving across, and in Sweden there's especially one political party that has a very clear view on what they think about digitisation in in school, in schools particularly. And unfortunately, it's a very uninformed position that they have, but it's it's a populist position. And in Swedish their position also rhymes, and so it makes it very, very sticky.

Henrik Göthberg 3:33

But I but I I think then to look at this objectively. Okay, we have political parties, and with our political parties, we have different views and opinions, and we're allowed to have that. So, what we are seeing now is the balancing act of politics and making something work and agree upon, where we come into the topic from slightly different views. Then me and you would argue potentially their populist views and they are uninformed views, they wouldn't agree with that. But that's essentially what we're seeing happening. That in politics, which this is, in order to sort out where do we spend our money or focus on AI, is in it's a balancing act that is also not driven by pure, how should I say, scientific yeah? I mean, I'll I'll understand. I'll cut you in here as well.

Anders Enström 4:22

Sorry for I was just thinking about that. I think that AI can democratize the whole system, you know, for each student because now the students can have like a tutor at home that is gonna explain to them. Uh, so it's not depending on if they are rich or poor. And I think that closing out AI from the strategy, from the national strategy. I think that's a big mistake. But I think the party that Henrik is talking about. Uh should we say the name? No. We don't need to go there.

Henrik Göthberg 4:52

We can look at this objectively, and then so from from anyone who's deep into this, it feels like an uninformed view because they had it's not thought through the whole way.

Anders Enström 5:02

I guess yeah, I think they are painting the floor and sitting in the corner now because they have a problem getting out of the corner because the color is still wet. Uh, so I think that they need to they need to have some advice from someone else than their uh spin doctors.

Henrik Appert 5:16

I also think I also I I completely agree with you, and I think you know they also dug deep into this position before the large language models started making uh their development.

Henrik Göthberg 5:32

So I mean like politicians, we have to be able to everybody, I mean like the methodical doubt, the scientific approach highlights this is maybe one of the most important traits of humankind to understand when your world model is wrong. You know, all models are wrong, some are useful. Right. It's your job to understand when you when your worldview or model is inherently so wrong, so it's bad for you. This is George Box, a famous statistician, made a beautiful quote on this. But and and then Adam Grant writes a book on on methodical depth, you know. Think again. This is what we're talking about that we we need to be very good at letting go of our dogma, and this is one of those things that they have an idea that might even be more merit in that before LLMs really came to the table, and now we we now start seeing deeper what this can read. So to change your position needs to be allowed, I think.

Henrik Appert 6:28

So I to me, the specific issue at hand here um is really not even related to AI, but it's more about uh the uh the lack of nuance nuance um of the the discussion that we are having. Like um they because of like their position basically equates all use of digital technology or or screen time, if you will, as the same, and that is just simply not so. Like uh, and I'm the first person to think that students should spend a great deal amount of time using their physical bodies, spending time out in nature, working with uh physical manipulatives and understanding how much does a kilo weigh, and uh how long is a meter, and so on and so forth. And I don't think that students, regardless if they're in you know kindergarten or high school, should spend klass lesson times on TikTok or Instagram or like dead scrolling or something like that.

Henrik Göthberg 7:40

So we can definitely do digital verra wrong if we're not doing exactly but if you think about it, the how to balance and do it right, is what we're talking about here then.

Henrik Appert 7:48

But but what they are not talking about is like all of the advantages that can come from using really any type of technology in the type or in the way that it's intended. So, like when people say we shouldn't use this technology like in schools, it's like well, any technology can be used in a good way or in a bad way. Like a pen and paper is also technology.

Henrik Göthberg 8:12

But but I think, isn't this then the the root issue here in communication and politics is that the question is complex and nuanced. But I don't think it's that complex. I think it's unintelligible. Like I think that they have But But hear me out. And what I'm saying is, what I see is that people are making it shallower and shallower and simpler and simpler to making almost all conversations bipolar, to be honest. There is there is no it's like it's like the extremes we take out in our in our arguments in order to position ourselves differently to the others, and it's I mean like this is what is happening in American society quite a lot right now, like how communication works that is quite hurtful because we are not nuanced. I think it's easy as well as you are. I think it's a no-brainer, but uh but I'm reflecting on what is the mechanisms at at work here. Yeah, I see what you're saying. And and I think we are we are now victims of one of our own short attention span. Of our own short attention spans and our own uh we want to make it really crisp, but we our communication is all about blah blah blah.

Henrik Appert 9:23

Yeah, and then we end up in the so if it rhymes, then it has to be true, right?

Meet The Guest And Co-Host

Henrik Göthberg 9:27

It has to be true, right? Yes, and okay, we're gonna have a great podcast today. Well, uh what a good uh Yeah, I'm excited. This is fun, true, yeah. And so with this, I really want to welcome both now Henrik Uppert as our main guest, and now we have a guest host in Anders Enström.

Anders Enström 9:49

Yes, thank you. Thank you for having me.

Henrik Göthberg 9:51

And uh I mean, like the the theme I I think is uh let's put the theme on the table because what we're gonna talk about today ultimately is AI and the future of mathematical thinking and learning. So we can start in how we learn math and how we can do that better, but I think it's a bigger question on mathematical thinking and thinking in general, especially in science, engineering, and stuff like that. Yeah, well, I think therefore I am, right? So, what about the mathematical angle on this, right? And just you know, you Henrik has been doing a lot in this field with magma math and and of course uh all that, and and you will get introduced uh mainly now very fast, but I want to actually start with you, Anders, because uh so we can leave you and go back. Okay, Henrik, thank you. But Anders, you've been on this podcast before. Was that two years ago? I think so.

Anders Enström 10:45

Yeah, around two years ago.

Henrik Göthberg 10:46

Yeah, and and my reason for inviting you as a co-host today was because I think um in you know the discussion we had then and the reason and this storyline where you have been working in gymnasium in in secondary upper school and really innovating and experiencing this. You to me are one of the real experts. Oh you know, I Andesh is a Andersh is a global superstar. I I I you're a global superstar, you're associated what is it, Apple, you're an Apple advocate in terms of a distinguished educator. Yeah, so it's the point is this you're you're a distinguished educator, which is a very, very hard position to get from Apple because they are looking at how we innovate education with technology, and of course, Apple wants to sell Apple, but they are actually wanting to sell better, wanting us to do better education. So that's cool. So could you just give us uh the the you know the one-minute version of of uh uh uh of you, so to speak, and then we're gonna move over to Henrik.

Anders Enström 11:50

Oh my god, teacher since 25 years uh back, and I have been teaching math and science. Henrik has been visiting my classroom as well, uh with a bunch of guests from US, I think. Yeah, yeah.

Henrik Appert 12:03

Uh when did you guys meet first time? Oh, has to be some conference back in 2019, maybe or something like that. Maybe set.

Anders Enström 12:11

Yeah, maybe 1819. Yeah, yeah.

Henrik Göthberg 12:13

So you used to paint the picture that this is a good mix we're having.

Anders Enström 12:17

Yes, it is, and now I'm holding lectures pretty much around the world. Yeah. Around AI, uh digital tools, um, talking about magma sometimes when it when it fits.

Henrik Göthberg 12:28

But ultimately, it's about how do we innovate the learning for children from which ages?

Anders Enström 12:34

Yeah, I have some some saying that how do we make school even more fun than it is today? How can we make the students long for Tuesday when they go home, go home on Monday? So so that that pretty much is uh what I'm talking about.

Henrik Appert 12:47

Yes, yeah, okay. And then I mean Anders, you have some pretty amazing outcomes with your students as well.

Anders Enström 12:53

So I just think it's worth like a minute uh talking about uh you know the impact that you've had over on a lot of lives on students and I hope I hope that I have an impact on my students' lives, and I think that that it turns out when they are reaching out to me after they have left me for for another school or another education system, and when they write wrote to me in on Instagram or a text message, and uh I get really really proud.

Henrik Appert 13:19

But it's not that I can We endorse that type of Instagram use in this in this podcast.

Anders Enström 13:25

Talking about social media, social media when you're talking about your teacher is it's okay.

Henrik Appert 13:31

Send your teacher some appreciation on Instagram.

Anders Enström 13:34

I think so. But I think that but I think that when students really like to go to school, when they feel that they are learning, and when they feel that I as a teacher believe in them, I think school is much more fun. And if they can and if if a digital tool can help us to you know, like prioritize right in the classroom so you don't have to do what you already can, but you can train on whatever you need to be more capable of doing and the skills that you need, I think school is um has taken the right step, uh step forward.

From Facial Recognition To Magma

Henrik Göthberg 14:06

Yeah, so you you you are a very, very qualified co-host today, Andre. Thank you so much. Thank you. So now we turn our attention to you, Henrik. Henrik Apert, founder and CEO of Magma Math. You know, can you do a little bit about your origin story into this? Like, who are you? Like and how did you sort of get into how did you fall into AI and into the problem that made you found a company? Like your background story to explain.

Henrik Appert 14:33

Yeah. So um I'm Swedish, I'm from Gothenburg, I have a background in in engineering. I worked a little bit in you know private equity and wealth management, and realized that that's it doesn't get me excited. I was having the the the Sunday scares, like feeling anxious about going to work, and so I felt like I want to do something uh that excites me. Like life's too short to spend it uh uh working with something that you don't really enjoy. And I've always had a very entrepreneurial mindset, um, been interested in technology and also interested in in math education. And so back in 2015 there came a PISA survey which showed that we had very poor results in uh mathematics, not just in Sweden but uh around the world.

Henrik Göthberg 15:28

Yeah. Is it mathematical STEM or is it most it's purely mathematical, the PISA survey?

Henrik Appert 15:33

Uh so PISA looks at a lot of different uh academics, but it was particularly mathematics that I was like thinking a lot about, and it was particularly the math results that were um that were you know how old were you to 2015? Can I ask you? Uh so you're asking me to do some math. So uh I was about uh 26-27 years old uh at that point. No, but I but but uh you know if you want to go back a little earlier, like um I founded my first AI company back in 2013, which then pivoted and became Magma in the story in 2015. So I was previously working in the facial recognition space, and in 2015 Facebook released something that was very much along the lines uh around what we were working around, and so we realized that there's no point in throwing good money after bad money, and so we asked our investors either you can have the money back or we have this new idea and it's about how do we fix the huge So you were you doing that at the same time as you had it had this in in quotation mark the day job?

Henrik Göthberg 16:43

Uh like yes and no.

Math Failure Rates And Equity

Henrik Appert 16:45

Um so um like I actually founded the company when I was earning my um my master's degree. Uh so I was doing it, you know, during my master's studies, final years, and then uh to make ends meet now and then I I had the chance to work a little extra uh here and there, but it it's always been like my main occupation since uh since 2013. All right, cool. And so, and so so that kind of um so to just finish how I how how I ended up here was we realized that there's this huge need uh uh in in math education. And in my background as an engineer, we we talk a lot about like how do good systems uh work. And like all modern organizations base their uh decision making on data, regardless if you are manufacturing cars or if you are selling toothbrushes on Instagram, uh, you base your decisions on data. And that needs to be true in how we develop math education as well. And it's particularly true in math education that we know that students base their new knowledge on their existing understanding of the subject at hand. Uh, and so in order for math teachers' lessons to be meaningful and efficient, they need to understand well, how are my students currently thinking about and understanding the math? So I can meet them at the point they're at. At the point that they're at. Like what is the unlock for them? And what ends up happening when that doesn't happen is that teachers progress at a faster rate than the students are tagging along with. And when that happens week over week, month over month, year over year, at the end of the day, for the students, it feels like the teacher is speaking a foreign language and it drives disengagement. And they tell themselves that they're not good at math, they're not mathematicians, they're not smart, and it's all completely unnecessary.

Henrik Göthberg 18:40

But this is so hitting my hitting home. I have three boys 19, 16, oh shit. Oh something, something, something, 19, 17, 15, I mean. And of course, uh each individual having different ways of uh you know succeeding and organizing in school. And the way you explained it now, about they are out of sync where the teacher is at, I think that's especially in math.

Anders Enström 19:09

Yeah, I I keep on thinking about how true that is. And math have some kind of status in in education system of being the most important subject. Uh I met a lot of math teachers that are, you know, like walking around in the hallway. You know, I mean I'm a math teacher, so yeah. But it's a it's a it's a subject as all subjects are, but you need to find the trigger in each student, and I think that data analysis could be one of the reasons or one of the purpose to get there, because it's so much easier for the teachers to get there.

Henrik Göthberg 19:47

Henrik, you you call it uh the it's a language. I mean, like if you treat if you think about it as a language, it is a language, of course it is a language, but if you think about it like you go you go to uh you know you go to Spain, you go whatever. I might know some Spanish, right? But if if a local guy shoots off in 120 million uh you know words per second, of course you you you you can't follow, right? So um I mean so it's so that so to to clarify, math is a language, math is a way of thinking, and therefore we need to make sure that the way we are speaking about it is at the level that you can speak that language. Right. It makes total sense.

Henrik Appert 20:25

Yes. And I mean I would argue it's a language that you know it's the most by definition, math is a logical language.

Henrik Göthberg 20:32

It's a universal language because it is the laws of physics, it is the way we are built up, blah blah blah science.

Henrik Appert 20:39

Exactly.

Henrik Göthberg 20:40

So it becomes very fundamental if you're not speaking the most universal language.

Henrik Appert 20:43

And and you know, I would argue that it it develops our uh logical and critical thinking when we develop uh our mathematical thinking and our ability to communicate mathematically. So if we're striving towards um a population or uh as a country that we want you, we want our citizens to be informed, to be able to participate in the uh dialogue to make make decisions for themselves as well as for society. If we want them to um be able to distinguish between misinformation, like these are all muscles that we train when we train uh when when when we improve our mathematical thinking. You know, Abraham Lincoln went back and he was a lawyer before he you know became a president and he went back and studied Euclid to improve his law like practice for his um arguing.

The System Problem Behind Falling Behind

Henrik Göthberg 21:39

So I hope people are listening are sensing now that oops, this is quite fundamental now because we are now talking on an AI podcast where we're talking about thinking, reasoning, AGI moving forward, yeah. And ultimately, now how do we make the AI safe to do that? And and what what is then thinking and logical reasoning? Where then ultimately, when you follow that consequence chain, understanding math has something to do with understanding reasoning and logical thinking, it becomes huge. I want to move into the uh the the question what was the main first problem that you I mean, like you're almost uh going into we need data on the understanding where we are, but was that how you shaped the first problem statement when you found it? You know, what was the original problem you're trying to solve uh with magma?

Henrik Appert 22:30

Yeah, so I mean, literally the problem was so like the PISA report was the wake-up call, and then in Sweden we have something called national standardized exams. Uh, in we have them in year three, six, and nine, and uh we have them in three core subjects, uh English, Swedish, and uh mathematics. And uh the fail rate in uh English and Swedish are between three and five percent. Uh there's a subset when we talk about students who have Swedish as a second language, but like let's not get into those. Yeah. But you know, for for the vast majority, the fail rates in Swedish and English are between three and five percent. Yes. And in mathematics, when we started, it was 18%. So almost one out of five students not achieving a passing grade uh in in mathematics. That's huge. It's huge. And you know, when you think about what the what the math or what the like the education costs, like we I'm proud that the Swedish system invests in in the education system, but we invest roughly a million crowns uh just in the education. There's other costs as well per per student. And you know, if if we think about uh K through nine education as a product kindergarten through ninth year.

Henrik Göthberg 23:44

So basically third through nine is kindergarten through ninth year. Yeah, so grundskola. Grundschule in Swedish. Grundschola in Swedish. That's a new expression for me. I'm not I'm not the nerd.

Anders Enström 23:56

We will do the nerd thing.

Henrik Göthberg 23:57

So I love you nerding on school stuff.

Henrik Appert 23:59

Yeah, I can nerd about this all day.

Henrik Göthberg 24:01

How many hours do we have? You know what? We we had podcasts running three hours. Okay. Two hours is a good let's see what we put our listeners through. We've done it, we've done it in the past.

Henrik Appert 24:13

But uh, but anyways, so um um uh uh K through nine education costs roughly a million crowns per student. And like if you think about this as a product and imagine a silly factory where you are producing a product and every single product that you like that comes off the assembly line, and every single product costs one million crowns to build, and one out of five products is is defective.

Henrik Göthberg 24:44

This yield, now I'm gonna nerd with you. This yield in manufacturing would not be acceptable. Like if you were if you were the factory manager.

Henrik Appert 24:53

If you were the manager of that, you like you would be out of your job very, very quickly. And I don't understand how we have been like accepting this year after year after year. And the thing is, like, it's a whole lot worse because these are individuals that we are talking about. We are not giving these students the start in life that they deserve. Math is the single best predictor for future academic and professional success. So if we are serious about providing all students with a start in life that they have a right to and that they are capable of, we need to get math education done right. And it's ultimately a question of equity.

Henrik Göthberg 25:30

But it's can you can sorry, uh, you you go. No, no, no. No, because my question was a little bit like if I'm if I'm very, very cynical, like Scania has been one of my biggest uh customers for years manufacturing. And if if we had a problem in Scania and we look at the manufacturing, and let's say uh we have uh three three to five percent yield in the gearboxes we build, and we have uh 80% yield, but like you know, can't use it, can't send it to customers, throw it away, but in the engine. I mean, like we would look at the fundamental yield in how we understand our business, red flag, or as Scottiness would say, golden arrow, to fix the things at least to the level of the rest. You know, and in in what I'm also thinking about in this context that oh, we should do AI everywhere and blah blah blah, but just looking at the data, the 18% failure rate in math, and this is the actual fundamental three core subjects we are teaching. This is why it's so simple to put efforts here to get that to par with the rest, then we can talk about fine-tuning the system further.

Anders Enström 26:46

I think you do some kind of you know, question after each school year, why is it so and why is this happening? But I think that it's so easy to blame the students, like you know, oh, it's the student's fault, it's never in time. Uh, no, but it becomes a systemic.

Henrik Göthberg 27:01

I mean, like as exactly individual, you can have this argument and uh why one gearboss was fucked up, but if you look at the systemic view as as an assembly line. Yeah, when it's 18%, it's not the student's fault.

Henrik Appert 27:13

No, it's it's how we're doing it.

Henrik Göthberg 27:15

So this is then it is systemic, right?

Henrik Appert 27:17

It is systemic. Yeah, of course. So this was this was the analysis that we did. Like we saw that you know, math is really like if we fix that, so many more students are eligible for a higher education. We set them up for uh success in life. And what we realized was that during math classes, there is so much valuable data that is arising regarding how they are thinking about the math. But the problem is that all of this valuable data is stuck in the students' workbooks.

Anders Enström 27:45

Yes.

Henrik Appert 27:46

And so, like our and since I had been in the AI space previously, I realized, well, if we could get AI to be capturing the students' thinking, yeah, then we can both immediately give the students the support and feedback that they need, but we can also aggregate this data to the teacher or to the principal or to the head of the municipality or even to school market who is responsible. And we could be we can be aggregating this in real time so we shorten the feedback loop to whoever the uh you know person of who is interested, yeah, so that we're able to adjust the teaching. So that was basically like the idea. Now we also think that because you earlier talked about your your your son, your sons, and that they can be in different places and they can also have different styles of learning.

Henrik Göthberg 28:34

I think they have different styles. My point was they have different styles of learning, yeah. And and and and most interestingly, one of one of my sons, I would argue, have cracked the code, how to analyze the teacher, and and how to plan and have a uh not only have not only learning but having a system for himself how he learns. So so the biggest difference between my boys is not in in intelligence. I think maybe the guy who's doing the least good in school is probably the most intelligent, to be honest, but he hasn't cracked a a learning system for himself, whilst the other guy is super good at planning and understanding and setting a goal and breaking his goal down, so he's used to having less work to do because he's structuring himself better around it. So I'm I'm so so the difference I see with my with my um, I mean, like is like hmm, it has nothing to do with learning capacity, it's just about their the capacity to organize themselves and be disciplined around that, I guess. Does that make sense? I don't know. Sure.

Henrik Appert 29:39

I mean, and I think there are a lot of different things that you know impact students' academic success and further success in life.

Anders Enström 29:48

I think like both but I think that I think the school could teach those strategies.

Henrik Göthberg 29:52

Yeah, we should those strategies is part of teaching, also, right?

Anders Enström 29:56

Exactly. So I think we need to and uh we in some how we need to make uh different kinds of strategies visible for the students, yeah. So I don't think that one size fits all, that's pretty that's pretty much what school is today.

Henrik Göthberg 30:12

One size fits all, yeah, and it shouldn't be and it shouldn't be. But there's almost like a big pop pocket of we can come back to when you look at this, how we can do the one-to-one view of this. Yeah, but I I before we get there, I I I really want to so if you now take that sort of problem statement and the vision statement, so how would you frame what the product is doing and what it's for? Because I can see both the understanding for the student itself, but it's even much more critical that it's a it's a tool for the teacher and even for the teachers' bosses and higher up. So so could you could you may could you help me understand what from this problem statement? Oh, we this is something cool to solve. What is the envision product? And and maybe you can tell it more from the context of today rather than we started on as humanity. But the truth drone.

Henrik Appert 31:00

It has, but it you know, the the the entire idea and concept really hasn't changed.

How Magma Captures Student Thinking

Henrik Göthberg 31:04

Like it's actually the core idea is understood from the beginning, but what is the feature set where you how you feel like it's a good idea? Yeah, I'll explain it. Yeah, okay.

Henrik Appert 31:11

So um, so so the final like so we have this conviction about you know, students they base their um their their new knowledge on their existing understanding, that piece I told you about. Yes. The other second fundamental belief that we have is that in mathematics, the most important thing is not if you get the answer right or wrong, but it's how you're actually thinking about the mathematics. How do you arrive at the answer? Uh that's why teachers say things like explain your reasoning, show me your thinking.

Henrik Göthberg 31:41

Yeah, yeah, yeah, for sure.

Henrik Appert 31:43

Um and so when we looked at what was available in the market before we started building magma, we saw that every single solution was just either multiple choice, A, B, C, D, or type in the answer with the keyboard. So, what we've done, what makes magma unique, is that uh students instead of doing their work in a um textbook or on a piece of paper, they get the problem presented to themselves on the Magma platform. It could be on a website, it could be on an iPad, it could be on a Chromebook, any type of digital device.

Anders Enström 32:19

And you can have it read out loud as well.

Henrik Appert 32:21

There's so many accessibility features, you know, access it in 140 different languages, a bunch of like uh bells and whistles. But uh the the core concept of the product is that students get the problem and it can be aligned to the curriculum, it can come from, you know, we have partnerships with curriculums providers. They get the problem and then they solve the problem, still doing math by hand, because we think mathematics is creative. So they are showing their thinking. Digitally on a pad or on an iPad, or they are so your iPad, you're seeing how they're scribbling literally. Exactly. Yes, and we analyze that with AI in real time, so we can give the students feedback, and then everything that the students are doing is streamed to the teacher in real time, and they can get this data in a lot of different ways. Uh, but uh, for instance, one view we call it a heat map. There you have all of your students, you have the problems, and the teacher can immediately see okay, here's a student that's quite struggling with the concept that we're working with right now, uh, or here's a problem that several students are struggling with. And they can also see how did all of my students actually solve the problem, so that we can have math lessons that are much more collaborative, where we can drive mathematical discourse, which is based not on the thinking of what it says in the textbook, or of how the teacher, maybe Anders, is thinking about it, but we can talk about how are the students talking about it really. And then we can also aggregate this data up to the municipality. Uh, so we can see, oh my God, this school over here is really doing a whole lot better than our other schools in algebra in year eight. Well, that's so interesting.

Henrik Göthberg 33:59

Yeah.

Henrik Appert 33:59

What can we learn from that teacher?

Henrik Göthberg 34:01

What can we learn from that? This this sounds like science fiction to me compared to you know, when you're walking into the normal class environment, and this already exists today. Yeah, yeah. So it's but is it, I mean, like do you see what I mean? I mean, like it's so because if you have a data-driven approach, and we're gonna go to the later because it it completely flips the teacher's role, of course. You know, let's let's hear how Anders talk about it.

Anders Enström 34:27

It flips the teacher's role, but it also flips the time in the classroom because I'm saving a lot of time just by having Magma analyse the tasks or the assignments, and I can walk to the students and you know, like talk to them and get a good relationship with them and talk about other things than just learning or math in that case.

Henrik Göthberg 34:48

Because um, of course, this is a lot of time. This is such a standard AI problem. If you oh, we want to AI sugarcoat an analog process, it doesn't work, right? No, you need you need to take the data, the AID, and reinvent the workflow. We talk about this all the time, regardless of role. Yeah, yeah. So, this is this is such an interesting thing, right? Because what you need to I flipping is the wrong word, you need to reinvent the workflow around math and maybe even adjust your objective function of how you teach.

Anders Enström 35:21

Yeah, yes, I yes, I think so. But but I think that we with AI in schools, I think we need to, you know, like flip the whole thinking about what learning is today. Because when I started 25 years ago, I could send home an assignment on Tuesday and say this is gonna be in by Friday, and you're gonna talk about in your text, you're gonna talk about what's the difference between um I don't know, something like that. And the students got home and written and handed it in to me on Friday. But if I give that task today, I can see 10 students going to the bus, you know, like open up ChatGPT, and also and then I can hear it in my app as bling bling bling bling bling bling bling, and the and the assignment is already in. So I think we need to approach learning in a different kind of way now. And I think that what Magma has done is that they have Henrik is talking about the the big picture, and I can maybe talk about you know how I've been using it in the classroom.

Henrik Appert 36:17

Yeah. How long have you been using it for, Anders?

Anders Enström 36:19

Just helped uh five, six years.

Henrik Appert 36:22

Yeah, so you've been one of our one of our like from the early days and then seen the development of it.

Henrik Göthberg 36:27

Um by the way, it was funny. I invited you as a co-host because I thought you would have some maybe you have some opinions on on on this. You bet you that was my intuition was right. I didn't know Henrik, it wasn't planned. I should have said it was planned. Yeah, you can say it on good research done by Henrik Gjöttberry here.

Anders Enström 36:50

But I think that you know, like like to be able to personalize each and every lesson for each and every student, for each and every week, each and every month, each and every year. That's the big strength. Because Henrik could be really good at algebra, but has some problem with geometry. So you need more assignments on geometry and work with that. And I can assign that to him. And I think that's the I think that's the most important thing that school needs to be each and every student.

Henrik Göthberg 37:22

Then it starts. I have the same conversations regardless of what enterprise type customer working on. It then you need to, as a teacher in the classroom, you need to have an objective function and an idea and what you want to do that goes from one size fits all to first of all, you need to ask yourself the problem is how do I do one-to-one uh plans for my students. So if you don't, if you're not willing to go here, then the whole thing falls apart, in my opinion.

Henrik Appert 37:54

I mean, I would say yes and no. Like I think technology can also support that. And then I would also argue the case that I think the future of learning is both uh individualized and also collaborative. Yeah. Like I think one of the main strengths of uh the education system where we have students commune together, you know, they come from different backgrounds, they come from different families, they have different cultures and experiences and all of those things. And then school becomes a melting pot where they get to meet one another. And ultimately, when they are having you know a good math lesson to me, they are not just siloed solving math problems specifically for them, but they are they are comparing and they are debating and they are thinking, what about if you solve the problem this way? Or I was thinking about it that way. Right. And so and so, how do you get those types of math lessons? And so to me, I think you know, meaningful implementation of technology, regardless if it's a pen or paper, or a whiteboard, or an iPad. It doesn't matter. Like, you know, but we should use technology to drive the type of pedagogy in classrooms that that is proven to have the outcomes that we want.

Henrik Göthberg 39:11

Well, because what you did now was exactly going all the way back to going from a naive bipolar conversation, oh, one size fits all, individualized learning. This is me who is not deep in the topic, who are simplifying the the logic. And everyone who's outside of school, so you're in good company, Henrik. You're wrong, but you're in good company. This is so good because it explains the political problem here and everything. Because what happens now, I I I this is oh, what a cool uh insight for me. This was you are I'm I'm seeing it from a um how do we re-en how do we re-engineer the working experience for the teacher and the working experience for the student? And we can have some ideas that we need to basically go uh one size fits all, is the simple that's that's that's the slogan, right? Or uh one-to-one for you know, that's a different slogan. And in reality, you are highlighting that actually we want to loop learn as a group, the sacred sacredness of the classroom we will talk about later, but at the same time, we need to recognize the individuals and their levels in this. So, this now to me, that's where we need to get to in understanding to really solve the right problem. Yeah, I think am I getting it right? I think so, yeah.

Henrik Appert 40:34

Or no, I no, I like I agree, and and I think you know how we should think about like in our case, we are not trying to replace teachers in any way. And you were talking about that earlier, right? Yeah, we're like we we we think about teachers as the most valuable important person, important uh resource in the classroom. Um and then but some like if you if you just break it down and put all of the activities that a teacher does during a day during during a class, there are those activities that add value, and then there are those activities that they have to do, uh, but don't really add value. And to me, value adding activities for a teacher are things that are advancing their students' uh like emotional or intellectual intelligence, comfort, uh, curiosity. These are things that teachers should be spending their time on, and they shouldn't be spending their time standing by the copying machine, printing up 30 different worksheets or grading things that we can use technology for.

What AI Actually Does In Practice

Henrik Göthberg 41:46

Could we do now a small sidestep? But it's not a nice sidestep, but but could we now peel the onion one layer and and talk a little bit about what is the AI dimension in this? What can AI what makes this not doable without AI or data? Like if you if you go into a core system and where team if we didn't have AI to understand what they are scribbling, we couldn't know. What is the what is what is that AI unlocks in this perspective for Gramma?

Henrik Appert 42:16

So again, like we've been built on AI since we founded the company back in 2015, and then of course, like with more advanced models, we can do more advanced things, but it's everything from handwriting recognition.

Henrik Göthberg 42:26

Okay, so hand so handwriting recognition is a core feature that's a core feature that you have to figure out. Okay, what are those things that are really AI driven? Because it's also about making a great thinking and system and product, yes, yes, yes. But what are the core AI features in this that we need to have that this cannot really be solved without analytics or AI?

Henrik Appert 42:45

Yeah, so again, like handwriting recognition is one of these things. There's text-to-speech, there's speech to text, uh, there's uh just like um different language, uh exactly language like translation um features.

Henrik Göthberg 43:02

So okay, so this is AI sort of machine learning stuff, and then we have the stuff that we're used to visualizing data in very smart ways, which is very powerful on its own, right? Sure. Like the heat map and all that. I mean, like so how do I how do I how do I make an insight pop? Right. Right. You know, in a table in Excel, I don't think it pops there.

Henrik Appert 43:22

No, I think it's stuff like this. No, but I think it's a great point, and it kind of you know, it it drives to the heart of one of the key things that we think about when we develop our product because our our hero users are uh teachers. Uh we care deeply about the student's experience as well, but really teachers are the hero user uh at Magma. And so, how do we make Magma really easy for teachers to use to have great math lessons with their students? Yeah. And um, just like students are on a broad spectrum of how they currently understand mathematics, teachers are on a broad spectrum. of their techn of their technikal, digital, AI um current understanding or abilities. Yeah, skill exactly. So we try to like, how do we make, to your point, how do we make the the core insights or the the like turning point insights, how do we highlight those and make them actionable? Because you can, you know, you can you can drown in data. Yeah. And the point is really well what what is the like one or two things that we really want a teacher to realize in this particular moment?

Henrik Göthberg 44:39

My word of we want the insight and action to pop, right?

Henrik Appert 44:43

Exactly. So we want it to pop and we want to help them like well how do you what do you do with this? Like here's the insight.

Henrik Göthberg 44:50

Is the insight pop? What should I do with this insight now?

Henrik Appert 44:53

Yeah what is the like what is the remediation? What do I do? Do I do I do I talk to an individual student? Do I go and differentiate my instruction? Do I um do I create a small group together with another student who has thought about this in a different way and I put them together and talk about it. Like there are so many pedagogical strategies. And so to be like to bring it back to you know when we're talking about the AI commission and how they're leaving school out of it. You know I think if you look deep at what it is that we're suggesting here, it's it makes a lot of pedagogical sense. And they wouldn't argue with that.

Henrik Göthberg 45:28

And and give me who has not used the tool at all so can you give me a picture in my head what what is the essence of the student experience of working in such a tool or such an environment and what is the essence of the workflow or the teacher's experience?

Henrik Appert 45:46

Sure. So a teacher can create an assignment they have access to material for all different grade levels so that they can differentiate and meet their students where they're at these can both be problems that they've written themselves. It can be problems from the magma database it can be problems from a curriculum that they have purchased or is available for free.

Anders Enström 46:08

Standardized exams is a good example.

Henrik Appert 46:12

Another great example standardized exams previous like national standardized exams and so they will create an assignment and this assignment is available for their students and it can be one single problem or it could be 50 problems like it's really up to the teacher so we put the teacher in control and the student will log on to their platform and they can see the problem they can show their thinking they can use AI to get support to solve the problem it understands their thinking so it can highlight to them where they are making mistakes.

Henrik Göthberg 46:44

Alright so here now we we get also a lot into generative AI and understanding. Yes because we you be you could because you get you get into the thing are we we all saw the first when someone used chat GPT I mean like that went viral and how he he's talking with with with the with LLM and he's based on how the the kid is responding he's guiding them without giving away the answer right so what you're talking about now this is a lot of AI because you're starting with okay we wanted to have handwritten because we want to have the real creative process captured. How are you thinking how are you scribbling your way down? Exactly and this is then feeding into your view of understanding where is he at in his thinking and where you can now start drip feeding the right type of nudges is that the way to understand until you until he feels happier that he solved it.

Henrik Appert 47:43

Yes and and we talk about so in in in mathematics we talk about um we we talk about um productive struggle so what is the right level of challenge that the students uh should be facing a fine-tuning from hell problem right you always need to be yeah like if it's not too big not to yeah if it's you always need to be right you need to have your nose just above the water yeah yeah good visual so and so to your point like how do we not give students answers but how do we ask meaningful questions to nudge them in the right way and you know this is part of how the system works.

Henrik Göthberg 48:20

Yes. All right yeah and I want to now move on with now I have a little bit better understanding on these topics there is there is a fundamental topic that you have talked about that I think we have already touched a little bit but it's a it's a it's a quite beautiful quote so I want us to unpack what we mean with the sacred dimension of the classroom. I think that's such a beautiful quote can we unpack that because typically you know when we get into politics oh we don't want digital because we want this to be very visceral we want to have the the the the the the human touch and all blah blah blah and I think you are saying the same but what do you what do you mean with the sacred dimension of the classroom? Because now we're getting also out of the relationship from of the one student and the teacher and into the classroom and the environment and the groups you know so what sorry for that long rant but what what do we mean with the sacred dimension of the classroom and I want to hear your story on that where it came from and I want to hear your perspective on it as well. Yeah yeah but let's Henrik start.

Henrik Appert 49:23

Yeah so I mean I I think I kind of touched upon it earlier. I think that uh you know we as human beings we are we are social uh animals social animals uh I think one of the core things that we learn in school uh aside from the subject is how do we how do we interact with one another like what's okay and what's not okay and uh how do I uh how do I get my point across to someone else and how do I also ensure that I'm capable of receiving information from someone else um and you know in I think I mean who knows where where the world is going to be with AI and everything that's that's that's happening. But one of the few things that I feel very confident in is that you know 15 years down the line students are still going to schools. Yes. Students are still going to a physical place where there is an adult who is responsible to cater to their learning experience both their academic learning experience as their and their emotional learning experience and I think that you know the number one and the number two resource for the student's learning experience is the teacher and it is their classmates.

Henrik Göthberg 50:37

Yeah.

Henrik Appert 50:37

So so to me it's like well they come together and they form a bond and they are on this quest, this journey together if you will to uh to ultimately like reach their full potential as human beings. And I think that you know I think it was Albert Einstein who said something along the lines of like the the eighth wonder of the world is compounding interest.

Henrik Göthberg 51:01

Compounding effects of different kinds.

Henrik Appert 51:03

Yeah but but I let me just explain where I'm going with that and then I'll pass to then I'll pass to Andrew pop Einstein in there without explaining what we're meaning. Oh it's not going to be it's not going to be that smart but like you know when you think about compound interest the most important factor about when you're compounding something is not that the rate that you're compounding but it's over how many periods are you compounding. And so I think that this is a beautiful argument for why we should invest so much more in our early education because the more we invest early in people's lives the more years we have to recuperate that investment.

Henrik Göthberg 51:41

This is your the the financial brain in you talking as well it's the mathematical brain it's the logical brain. But it's also start your pension saving early sure it's the same argument. But and how do you understand the sac dimension of the klass room? What spin do you put on that?

Anders Enström 52:00

I love that expression by the way it's so many dimensions in the klassrum. You know it's 30 students in one klass from different kind of homes and different kind of nationality and different kind of feelings and different kinds of well being so it's it's a huge range of dimensions in a classroom but I think that I think that we need to talk about more than we do today about students wellbeing and how students interact with other people or other students as well and I think that's a big part of it. When when I get a new class in grade seven I always tell them that's it's you and me against the world and the more and when we come to grade nine I say that now it's you and the world together. So it's kind of it's kind of a journey from grade seven to grade nine because the students when they are leaving my classroom I want them to feel that the world that they are stepping into or out to is their world they're gonna change this world into something better than it is right now. And I think that's the that's the hardest part of being a teacher you know like make sure that every dimension of a classroom gets their voice heard in the society in somehow. But I'm curious I'm curious Anders I mean you you had many cohorts of classes and to my understanding but you like I'd love to hear you expand on this like really every single class has its own yeah its own vibe its own vibe its own DNA would you just like just expand a little bit about that and how it affects the role as a teacher and and how you I how you work with that I try to be me all the way all day all week all month all year because I think I think the both the class benefits from that because they know that I'm me each time I see them and then then some kind of problems always turns up you know somebody's uh you know wants to sit at home or doesn't come to school and and it could depend on the group if the person wants to come back but I think that you need to have an open mind by solving problems every day. And I read somewhere that a student no a teacher takes 10,000 decisions each day and it could be big decisions or micro decisions. Yeah so I think that just being myself and being positive is the way to build that group that I want to have and I that I want to be with and we want to create our own vibe in in that classroom and Henrik has visited me a couple of times now I think and each time it's I have a lot of visitors in my class and each time when we get visitors in my class they give the the visitors an applause you know like welcome to to this classroom and I think that's a that's that's a cultural thing. We need to create a culture in the classroom where everyone can be themselves but everyone also sees the other person and knows that okay Henrik is a good person with a big heart and I want to work I want to try when I I I I got a sports metaphor in my head from that the sacred dimension of the classroom and maybe this is completely wrong.

Henrik Göthberg 55:36

So I want to test it on you and then and then you can kill it or approve it or uh confirm it. I'm not sure if it's true. But I'm thinking around this like if I take a sport like football I played a lot of football in my years and even more interesting when you look at elite football like Premier League or Champions League football where you sometimes see some some of the most expensive players obviously talented in their own right but they play absolutely shit. This is one angle another angle is that you know when you you get uh we you get uh a young team like like um now we're talking about um junior football or or boys football or girls or girls or girls football very important I I I wasn't then you get to the point where I can try to teach them exactly a system run this this this this this without having the creativity without having the understanding so so where I'm going with this metaphor is that I want to try on you is like we need to create the culture we need to create uh the greater than the whole it doesn't matter if I have the all the different all the different individuals they can get to one level on their own but they can get to a dream team status by by how they're working themselves and the really best teams they both manage the individuality of one player and and his greatness to be creative on the spot and do the uncertain but they also form that inside the system of the other actors in that team. So the greatness of the team is the balancing act between the individual's uniqueness and creativity or or you know what he brings to the table and what they can do with that together. And so the whole metaphor then also becomes the whole thing of the coach of the teacher how do I bring out the best of the team and how and how how do I how do I build a team i peer learning takes us further potentially than individual learning.

Henrik Appert 57:42

I love that metaphor I think I think it's it does it work I think it works.

Anders Enström 57:46

I mean you know Michael Jordan wouldn't have been Michael Jordan without the bulls right and the bulls and the Pippins and them and and what and if you if you that's a great story right what what what what is the contribution from the coach from Michael Jordan in that classroom in you know in in that gym does it work yeah yeah I think so because I have a saying when I was a teacher in my I have to count my stupid my son's age now but 15 years ago I always I was a teacher but I also trained a soccer football or soccer team in a small city small town just outside of Stockholm that's called Skugos and we were just a suburb gang from from Skogos playing soccer but we become a group you know we become a group there was there was some place that were better than others of course but we won the um the Poika Lsvanska 15 years old a small team we beat all the big ones you know like Roma Poikan Hammabi and whatever because they were taught that when you are here you're working hard and you have to be nice. So work hard be nice that's the my mantra in the classroom as well work hard and be nice and and the world is open.

Henrik Appert 59:04

Yeah I think it resonates and I think if we talk about I think I think if if we talk about growth and we talk about student growth but really growth in any in any dimension and I can connect it to to my entrepreneurship but I really think it goes for anything. I think that you know daring to fail yeah like and how do you create a circumstance where it's okay to be wrong and then fix it and pick it up. Yeah yeah I think exactly and I think I think this is the one of the the challenging things in mathematics as well because it's like since it is you know there often there is a right and wrong answer to a question. And so students who are can be very tough on themselves and if they get an answer wrong they're like I'm no good at math. And you know if we can foster a mentality in classrooms where being wrong is like this is this is a great opportunity and how do we learn from here and like how do we how do we how do we train grit? How do we like how do we develop perseverance?

Henrik Göthberg 1:00:00

How do we all this stuff you could take immediately and put that in a football team.

Henrik Appert 1:00:04

Exactly exactly and so I think it's it's one of the core core traits and core responsibilities of a teacher to like how do we create a safe environment especially to to try things out you need to dare to be a player yes but you need to dare to play.

Henrik Göthberg 1:00:19

Yeah if you're lockdown and I I don't dare to make that pass it doesn't work right and at the same time you will I mean like Jordan says it better than any way uh Federer says it better also you you know how many balls I mean like in tennis it's interesting you look at you look at the game they win right and then look at the stats of how many balls points yeah he won points he fucking won how many points he lost and the failure rate in each game yeah so Federer did this beautiful speech you can find it on YouTube yeah when it talks about his failure rate in each game part grit and and and and and psyche to to use win win because when you look at it that way he wins the game and it looks impressive but it but he wins around 80% of his games but he won like 55% of the points or 53% of the points. He wins 80% of his games but he he he hardly point wise it's not that big difference.

Henrik Appert 1:01:14

Yeah so he wins the points that matters and he treats every point as like this is this is the point that I need to win.

Henrik Göthberg 1:01:20

Yeah but it's interesting then because it means we need to be set up to deal with failure and we need to learn from failure as a group and as an individual.

Anders Enström 1:01:29

But I think there's there's a teacher tip for that because I've been using magma of course and you have a whiteboard in magma where you can you can project it on the screen and you can as a teacher write on it. And if you intentionally make some faults or if you you maybe you doesn't do it intentionally they became just wrong.

Adoption Is Hard And Why

Henrik Göthberg 1:01:50

And the students can find that you know they can say look look look you that's that's not correct okay I also make uh mistakes and and that's you use the technique yeah yeah because you know it's like uh it it's a teacher tip so everybody we love that use it we love that type of pedagogy like my it's called in English it's called my favorite no like what's a very common mistake how do we learn from it exactly and it drives engagement as well but yes it does all right so this is an AI podcast we've been far down the math classroom so so now let's now let's take it back to AI and technology and invention and let's let's take another angle on this right now um because we are not we have not been painting a picture of oh this could be great this is fantastic blah blah blah but in reality regardless if you're in the in in the classroom or if you're in public sector health or if you're in in in an enterprise the fundamental challenge is of course adoption and starting getting started from scratch and also understanding you know what is broken with the current way we're doing it sometimes we're arguing I don't want it I don't want this AI stuff la la la but we haven't even recognized what is broken with the current workflow whatever so so let's talk a little bit about um the the real context the real environment here you know in what way is the current way of teaching math not broken so we need to go in this direction we are we are painting a picture of a vision but I want to take a step back in understanding what's the context we're in now that we need to change from and why is this change hard?

Henrik Appert 1:03:30

So so so what so if I flip it if what the way the fundamental system is set up around mathematical education today and the way we are doing it in our classrooms on average we all want to be better but what what are the fundamental things that we are that we are making different when we're going in this direction now to where it is that's where I want to start do you want to go first Anders or should I I think you should go first because in my opinion this is a little bit like how do you understand the problem and and what is then well so I think change is change is always hard like by definition and it's one of the it's one of the like things that we think most about at Magma which is also like why we invest so heavily in ensuring that our partners and our partners that I'm talking about the municipalities in Swedish kommuner in the US it's school districts or trusts in the UK school boards in Canada Schultz in Germany like the markets that we're in now we we think a lot about how do we how do we support implementation like how do we how do we give teachers not just how do you how do you log in or how do you tool is not the point it's the workflow. How how do they get confident in this work experience? And and and I think again it's like here is where like leaders they need to create circumstances for their teachers to be learners that it's okay to to fail. And not step up and it's scary and we have these expectations of our students that they should be okay to learn and then when we look ourselves in the mirror it's like oh but I've been doing it like this for like 15 years and I'd rather not change.

Henrik Göthberg 1:05:13

But isn't that what so what you're talking about then is that we are set in our ways to to think and actually it starts with daring to look yourself in in the mirror and unlearn maybe what I've been thinking about as my mental model how to teach maybe didn't hit the mark on on on following the understanding is that what we're talking about here that people like teachers are not really understanding the importance of these fundamentals or not I don't know.

Henrik Appert 1:05:40

Well I think that they I think that like I mean first of all like I would never like talk down about teachers. I think teachers are phenomenal they do so incredibly much for you can do it I can just go but but it's another way of flipping it.

Henrik Göthberg 1:05:51

If you don't have any means of getting the data how would you track understanding well that is the problem which is like which is where we come in. Is that they don't have this data to act on. Exactly. So that's where we think you need to act on intuition. Yeah. Because to get to that granular level that you are now thinking we need to get to, how would you do that?

Henrik Appert 1:06:11

Yeah. I I can say that, you know, so in Sweden, we're now used in more or less every single municipality in some capacity, and about 65-70% of all Swedish K through 9 students have a subscription by their municipality. And in the last PISA surveys, it actually showed that Swedish students who are using a digital solution during math lesson outperform students who are not using a digital solution. So like Magma is now at the scale where we are proving that impact in surveys that are objective, which is really really cool. Exactly. It's not us saying Anders, tell us about the great results. You're happy to do that as well. So we are moving in the middle.

Henrik Göthberg 1:06:59

We are moving in the right way, but but what are we moving? What are we shifting out of in your way of the street?

Anders Enström 1:07:05

We are shifting out of tradition, I think. Because everyone has gone to school. And I think that's one of the biggest problems today. Students are logging in, and I'm logging in in in each uh start of the skol year. And I always get some kind of mail from a parent. Are they not gonna use the the regular math book? No, we are not. And then you have to explain. Because I can see what your child or my students are doing each lesson, and I can analyse that or AI can analyse that for me. And the students are gonna get those resources that they need to progress in their math education. And after that, it's it's uh it's a no-brainer for them. And they say just oh okay, let's go.

Henrik Göthberg 1:08:01

So let me put uh some wordings on this because this is my core business, what I'm working on when I'm working with uh with AI adoption in in core workflows. My core philosophy right now is in order to adopt new ways of working, whatever it is, there is there is some there is a fundamental process of unlearning. I mean, like you if you have a world model of how things should be, how I should teach math, you need to un you need to understand the blind spots in what you're doing. If you're not if it's not like we are not wanting to fix things, it's like it's a blind spot, I usually say. The gap is a blind spot. We haven't thought about it, we have we don't have the data, and therefore we need to then question the world model we have or teaching what we have. We need to unlearn what is inherently wrong in that model before we are ready to and understanding to embrace the change to find a better way to do it. And I think we're coming back to this fundamental problem over and over again. It doesn't matter which custom in, I I can observe behavior, I can observe how they are discussing. You you you observe the behavior of um uh the the teachers, oh sorry, um parents. Yeah, or you're not gonna have a okay. So you need to calmly not recognize oh, there's a past process here of unlearning, relearning, and then it's not a problem. No, and this unlearning, relearning, we skip all the time in the fundamental, you know, oh, we're gonna innovate with tech, but we haven't really made the arena ready to take on the tech.

Anders Enström 1:09:34

I think I think that that's the one of the biggest problems in ed tech in general. In ed tech. In general, I should say because because magma and other other services as well are really, really good for learning and or enhanced learning. But I think you know, to reach out to each and every teacher is the hardest part. Yeah, right. Because they need to change along the way, and that's like Henrik said in the beginning of this conversation: that change is a hard thing to make. Uh, so I think that we we need to have a structure around how should we do this in this case with with magma? How should we implement it? How should the students move on? How should the teachers learn? So I think you need to have some kind of way forward and a vision for math education today.

Henrik Göthberg 1:10:24

Yeah, but uh and to me now, we are touching a very sore spot or we're testing uh Pud and Shana. I I I've been chairman of the Data Innovation Summit for ten ten years now, and we be so all the vendors are there that to we talk about the technology and it's all fantastic. And over and over again, we we we we we talk about the tech and the invention and the product, what it is. But we we don't either you know we we we are failing to connect to the very simple fact that any type of tool that you bring into a workflow changes that workflow. It doesn't matter. I mean it should change the workflow.

Henrik Appert 1:11:09

Like it should be like if it doesn't, you then you're not using it intentionally, and it's when you don't do that, then so this is but this is such a simple, obvious fact.

Henrik Göthberg 1:11:19

But if you look at how people design change programs, training programs, we we we okay. Now I'm gonna educate you how you use uh magma. This is a wrong question. We need to understand what what is your workflow as a teacher and how will that workflow change when you use magma? I used the analogy back in the day, we changed into Power BI uh for controllers in Vattenfall, way back. And we decided now we're not gonna have a Power BI training, we're gonna have a uh business controller training according to Vattenfall standards using Power BI, which is obviously two different ways of looking at the problem of then getting the workflow for business controllers better. And I think I think the industry, we are failing miserably in this that we are we are we are looking at the invention, but this is Joseph Schumpeter, you know, 1940s, distinguishing between what's the invention versus the new doing. In innovation is to figure out the better way to do things the better process, yeah. So this is this is focusing on the teaching process. This is new doing, and then what can boost that way of doing things differently? That's the boost, that's the product. Yeah, yeah. So so we are we are poke we are talking about products, but products are only employed in order to do new doing better.

Henrik Appert 1:12:42

Yes, I mean, and this is really like one of our like sticking points has really been like you know, there's no intrinsic value to uh digitize for the sake of digitization. The value comes from when we improve our processes to get the outcomes that we want.

Henrik Göthberg 1:13:01

And so in Magma's case, like so we try to solve this in part by working really closely to our teachers, to be called to work with them to influence the way they're thinking about the workflow, but also for but also for them to influence us so that we design the product that fits exactly, yeah.

Henrik Appert 1:13:19

Like, and so like it is because to be a teacher is incredibly stressful. And if you are you know not too tech savvy and then someone puts a piece of technology, like Anders is super tech savvy, like probably more tech savvy than I am. But uh, so if someone puts a piece of technology in front of uh uh someone who's not very tech savvy and they don't know how to use it, and it's a stressful situation, you know, it it's very safe to go back to like let's turn to page 56 of the math book and keep doing the problems. Um so so we think about it that way. And where where we have immediate success with magma, there are certain things that magma does for the teacher that reduces work for them. So, for instance, they don't no longer need to go to the copying machine and uh you know copy stencils. They don't need to like relearn that. They they have no like natural inclination to like, now I'm gonna go and stand by the copying machine. So, like those things magma solves automatically. Yeah. Where we have where we, together with our partners, have additional work to do is well, how do we how do we support teachers in great, now you have all of this data, now you can see the students thinking. Instead of just you showing how to like solve the problems, how about we bring up emil's or Anna's or Mohammed's solution? Yes. And how about we let them talk about it instead? And so, like, that's where the like the reteaching to your point, how do we get to the next level? But I would argue that magma is all like even at the like just install, it creates a lot of value, but to get all of the value, they also need to relearn the.

Henrik Göthberg 1:14:58

I know I get goosebumps. So now I'm gonna restate the same question. The context is now the systematic unlearning and relearning and the machinery supporting the teacher and the education to figure out new doing and innovate with tech. If I take that so the the the core question is what is broken in the school system on municipality level, on innovation level, on how we run teaching versus IT, how we do tech projects versus you know, you know, business uh teaching innovation projects. You see you I'm painting a picture here, I'm guiding you. What is really broken?

Public Sector Incentives And IT Friction

Henrik Appert 1:15:44

So I mean that's that's a very complicated question, and then there's a lot of things that are like challenging, and it depends on like which like which system are you talking about here? But like I'll give you like one thing that I have found incredibly challenging in the public sector, yeah, is like um, and so when we founded magma in Sweden and we were trying to break into the Swedish market and get municipalities to say, like, yes, we want to um we want to pilot magma, we want to give it a go and see if it can have the impact that that we're saying that it will have, there was a lot of pushback, and you know, the municipalities were trying to push the decisions down to the principals, and the principals were trying to push the decision down to teachers, and we're saying, like, no, this is a systematic uh problem, and teachers, they it's like it can't be their job to also evaluate you know 70, 11 different products and make the decision of what's uh what's working. But so what I want but but I want to I'm gonna answer your question now.

Henrik Göthberg 1:16:47

And so like you could you already are answering it now. You have already the the first, if if I if I suck out a generic answer, what you said now, the fundamental decision flow, decision making, how that is organized between tech versus municipality on on strategic level versus operational level. That is this is here here. We have friction at least. I'm I'm not gonna 100%. You're like let's not say it's you could say it's broken, or I could say, oh, it's a phrase, it's a bit of friction. No, but definitely we have not organized these logical decision flows, so this becomes a meaning because you need a teacher's perspective on it, you need the municipality perspective on it, and you need to take perspective on it. And these ones are not coherently working towards one decision flow of how to select innovation to focus on.

Henrik Appert 1:17:35

Yeah. So but I want to go even deeper specifically on this, and like I think where the like where the core core issue is in the public sector, yeah, is that there is like there, like if you if you're working in in the public sector, like at a municipality, like if you take an initiative that is phenomenal and it just changes everything, there is very little like appreciation or like, oh my god, now you're gonna have a super career. So like there's very little incentive to take a risk on something new. If it fails to get exactly so, we have very we have very skewed uh like risk reward situations for people in the public sector. So they become guardians of just keeping the status quo.

Henrik Göthberg 1:18:24

Yeah.

Henrik Appert 1:18:24

And that is one of the reasons why we have so little innovation in the public uh systems. So I think we like it's again, we're it's a it we we come back to this. Well, how do we create an environment where it's okay to make mistakes? And make mistakes, exactly.

Henrik Göthberg 1:18:40

Yeah, so on all levels, on all levels, yeah. So here now we're talking about what what is inherently broken is you I can be technical on the decision flows, but it's just a systemic uh incentive way, uh the way we're incentivized, to be honest. Risk-reward balancing is you have all to lose and very little to gain. Exactly. Like so that I mean this is this is one problem of I mean, you have a very big question. And I can't wait. We are talking about roots versus symptoms, right? And we are now talking about two major roots. The way we are the way the decision making flows, the way it should be done, so it's coherent, is it it is not so. It is not without friction. And the fundamental model that rewards experimentation and learning uh it's the opposite, right? Sure.

Henrik Appert 1:19:28

And there are there are additional issues, like just competence is another. Uh, like like the world is moving at a very, very fast uh pace. Yeah. And it's tough for everyone uh to keep up.

Henrik Göthberg 1:19:40

Yeah, how how can you innovate something if you don't know the art of possible? So it's the balancing act. You need to you need to have your eyes on the invention, you need to have the insight on how that could change my new doing. And now you have to have competence and opinions about the workflows and how you improve them as what where's the friction as well as the technology that can solve that friction.

Anders Enström 1:20:02

Yeah, and I think when I'm out, you know, like holding lectures and meeting municipalities in school, I think that one of the biggest obstacles right now is the department of IT supports. It's interesting, right? Yeah, because they have always managed, you know, it's like a Windows computer, and I can all I that's the that's the only thing I can take care of. And then you come in with an iPad or magma or whatever tool that you are using, and they say, No, no, I can't I can't support that one. No, I don't know how to do it.

Henrik Göthberg 1:20:35

And this is back to the decision flow and understand the the facets of the executive decision, the technology aspects of the decision, the teaching aspects of the decision. And those now are different lenses on the same topic, but they are working in different departments with different agendas. Yeah, they don't have a common view on this the same problem we are dealing with in enterprise. Sure, yeah. It's the same fundamental problem here. This is what I mean with decision flow against exactly because they need to go, it needs to work 360.

Anders Enström 1:21:08

If it stops there, it doesn't get it over there. So yeah.

Henrik Göthberg 1:21:11

And typically failure is stalling, it stalls until the air runs out and then it's gone. Yes. They wait it out. All right, but but I I like the way that we could sort of quite simply highlight decision flow around these 360, many different lenses that need to work as one. We have the fundamental understanding of what was the second one you said?

Henrik Appert 1:21:33

Um I mean, creating a situation where it's safe to make mistakes and to take you like innovate and try things out.

Henrik Göthberg 1:21:40

So we need to go from a this is the example in enterprise. I ever we need to have a perfect business case, and if the business case fails, you you're out of there. No, no, no, we need to have an innovation cycle. Yes, exactly. And then this is that's the second one. And the third one is how do we systemically work with competences around this arena because it's fucking hard.

Henrik Appert 1:22:00

Yeah. I mean, but I would also to me, I also think specifically, like the fact like coming back to why we started, and when we look at the math results and seeing that so many students fail in mathematics, that in the like in that sense, I feel like it's been underinvested in. And then the initiatives that have been done have been more let's do more of the same. Yeah, like and just more of the same. Not rethinking. Like, doesn't like we're not gonna we're not gonna significantly move the needle if we just like, oh, but we're gonna do another math initiative, or we're gonna increase the teacher concentration from you know, yeah, 17.8 to 18.3, you know, uh, or the other way around, like uh 18.3 to 17.5 students per teacher. Like we need to like, this is a systematic problem. We need to change the system somehow.

Anders Enström 1:22:59

Yes, and they put put another math lesson in uh exactly. And it doesn't matter. Yeah, because the the structure of the lessons is the same.

Henrik Appert 1:23:07

If it's an uninformed math lesson, having more uninformed math lessons isn't gonna do it. No. So we want like, and I think if you go in, if you go into a classroom uh and it's a math lesson, which is where students are disengaged, regardless if it's with magma or or uh with with with a pen and paper or or a book or a teacher, disengaged students, like this is an example of like waste. Like we are wasting the students' time and we are wasting this the teacher's time. Yeah. And so then we need to change things. Yes. How do we make it interesting? How do we make it engaging? How do we get the students to like love math? Love math or love learning and just like think, like, you know, actively think. Like a mathematician. Yeah.

Henrik Göthberg 1:23:53

And so all of a sudden now we're bringing in another dimension of uh understanding change management behavior and motivations. That that that is also an ingredient that sort of when you when you build this as a technical implementation project, you're you're not you're not opening up those dimensions of the problem. So ultimately, you're trying to run a technical implementation where you have massive blind spots, you're set up for stalling or failure. You haven't you haven't sorted out the decision arena, you haven't sorted out the the reward arena, you haven't sorted out that the competences are there actually to take the 360 around the topic, and we haven't not really understood the the motivations and passions and that makes people want to do that. Those are kind of big topics that are kind of blind spots when we are over and over again staring at the tech and doing another IT project or out of it. I'm I'm rambling, ranting now. But do you see it? Like it it won't work if until we understand the real arena and the real story, what the change project is all about. Because it's a little bit like my view of this is like you have framed the project in such a way, so you are only looking at this. I usually say like this if you ask for a project scope, you get a project scope. Okay, but a project dies. So, how how do we make sure that we continue to evolve the life cycle? Oh, we need to have a product scope. So now you have shifted the whole lens where you're gonna ask your bosses for decisions around is it a is it a project scope? It's this mandate you get. Is it a product scope? It's this manage you get. And then I say, but the product is only there to be employed to use it in practices. So we need to ask for practice management scope when we build these investments in the project. And you and I think we are over and over again simplifying, and maybe the vendors or we in consulting is doing it to box something even simpler, you know, to get the deal. But for from an adoption rate perspective, we have not defined it as a practice management project. Or you see what I mean? It's like so we so what we are doing now, we are looking, we're adding topics to what this is really all about that I very seldom see defined in such a way when you look at it. Oh, we we're gonna use magma now, we're gonna have this implementation plan. I'm ranting, but I'm but I'm but I'm but but am I true or not? Am I am I off or am I maybe you are doing it better? But I'm seeing it in enterprise how we are always doing this tech mistake, tech-trigger mistake.

Anders Enström 1:26:24

Maybe maybe we do some mistakes now and then around EdTech. I think so anyway. Of course we are. Yeah, but I think that I think that politicians right now is not helping to implement ad tech as well, because they're going out and you know, like they're doing some bad rhymes and and and so and talk about that that we should ban the uh the screens from the classroom. And when they say that, both the head of municipality and the principals are gonna say, but the politicians are saying this now, and we can't really do that, you know. So it becomes like a like a stop sign for implementation of edtech at the moment, anyway.

Digital Divide And AI Access

Henrik Appert 1:27:06

I mean, where my fear is around that, and I mean I speak to leaders in municipalities more or less every single week and earlier days every single day. And uh I mean the fear is like when you're out there in reality, the leaders in the schools they don't want to go back. They want to like they see the benefits, they have the data to see the benefits from it. But the challenge is, you know, municipalities they have tough budgets and uh and uh digital infrastructure costs a lot of money. Yeah and then if you know when push comes to shove and budgets and they're thinking about okay, well, where can we cut? Well, there's a risk that you know they say, well, maybe we don't make digital devices available for kindergarten through three, third, third grade or sixth grade or something like that. And uh and and you scale. That back, and I mean now we're kind of like full circle to why I was at the event last week because that is really what I'm what I'm afraid of, and I'm afraid of it from an equity lens. Like I think that AI is this you know amazing technology, it holds a lot of promise, it holds a lot of risks as well, and we can get into those if you guys want, but it holds incredible promises for more efficiency, for more human creativity, for faster uh learning. And if we scale the system back, then there are going to be those students who still have access at home, and then there are gonna be those students who don't have access at home. And it creates a digital divide in society, and it's a competence divide. And we are talking about competence in arguably one of the most critical areas for future human uh flourishing in uh in in learning. And so why why me and so I wrote this this uh we we wrote a letter to to the to the government, me and a couple of other uh founders in the um in the tech space, AI space, and uh and and we wrote our concerns about this because we think it's really, really important that it's not just high school and upper middle school that have access to digital technologies and that their teachers have training for it. We think that all students and all teachers should have training in this so that when you have students coming in in in lower school or in sorry for saying lower school, but in in like earlier years or even middle school, and they have questions about these technologies, you know. Well, the best way we learn is well, we we we experiment, right? And so that needs to be ready to experiment with and try and test and and tap into the students' curiosity versus having teachers like having no idea and we can't touch it basically.

Henrik Göthberg 1:30:01

No, and they're here. Now we're talking about why we started this point. We're talking about the fundamental AI divide of the haves and have-nots that you can discuss on on you know macro geopolitical scale of the huge hyperscalers, but fundamentally, we're getting down to the AI divide on macro macro microeconomic society, how we how it becomes a very tricky society if there is a divide in society of the ones who is like a an upper or lower class based on competence in the most fundamental technology everybody will be using. Yes.

Henrik Appert 1:30:33

And in the Swedish in the Swedish skull law, skullagen, like one of the one of the like, I think it might even be the first sentence. It says that uh it should, it should like public education should level the playing field and ensure that all students uh get the opportunity to reach their full potential.

Henrik Göthberg 1:30:55

So it's like just based out of that, like well, how can the AI commission suggest something where we're like not creating creating equal opportunity for all students and but now let's tie this together now because now we have a fundamental that we need to learn AI and and and and and um uh at early age as one core topic here. And today we're having the topic of um mathematical thinking and learning. So now we get to the point where of course uh you know what? It's a pretty good idea to know math if you want to understand both AI and how to make AI safe and understanding the limitations of AI. So here now we have if you go back to the mathematics, you know, it's often described as the core language of science, engineering, technology. And if the if the AI is significant is now changing the way we understand how we learn mathematics or how we teach mathematics. So I so I just want to take the consequence chain, mathematics and AI learning in school, and to see how that really, you know, the consequence chain of of stopping the funnel of new talent here versus what happens as as us as a society when we can't sustain those uh other, you know, core jobs or core competences, science, engineering, technology. So could we look at the broader implication of math, math learning to do that well and what that sort of gives as a as the uh compounding network effects into having a more prosperous society like like these roles I'm thinking about.

Why Math Thinking Matters For AI

Henrik Appert 1:32:42

Sure. So I mean we can talk about so I think the question that you're asking is profound, but no, I don't think it's I don't think it's bad, but let's but let's if we can unpack it a little bit. Let me formulate the question I'm trying to add. Because I think I think that they are like we're both I kind of understand your question a little bit about well, well, how important is math education for the development and future of society, and also how important is math education and development for understanding of AI and how we develop this technology further. And so to me, like I think that math education is incredibly important, both just as a right for you know each and every one of us. Like we should all have a basic understanding of you know, if I take out a loan and I'm paying interest, well, what is interest and how does that work? And is like is 10% or 3%, is that better or worse for me?

Henrik Göthberg 1:33:39

And uh baseline you need to do in order to act in the normal society today. Yes, and so and so and if that society is full of AI, which is essentially math, what is then you know I I think you're getting it, uh you're on to it.

Henrik Appert 1:33:53

Yeah, well, so so I mean, so I think in general, like in we want to develop a higher understanding of mathematical uh in all of our society. We talked about it earlier, how it develops our uh critical and conceptual thinking, and you know we're gonna have a more complex world with more misinformation, disinformation, and we want students to develop critical thinking. And I think that mathematics is a great language to practice and train our communication skills, our critical thinking skills. So, with that lens, I think it's important for all. I also think that there's a small subset of people who are going to go in specifically to the mathematical field. Uh, and uh I think if you know, if you if you go back in history, then you've had you know these great mathematicians where they have been the ones doing the mathematical thinking and the and the developing of it. And I think when you think about the development of the mathematical uh realm and our understanding of that, I think that to a larger degree is going to, of course, be aided by uh by AI. And again, like AI is just like its computational power and it's statistics, and so we're going to like learn more about the field of mathematics and we're gonna learn more about all types of fields, we're gonna learn more about space.

Henrik Göthberg 1:35:15

But let me test this because I'm thinking what what we are seeing now, starting in software engineering or software coding, where basically we're getting more and more AI support to do the actual groundwork. So so and it shows a trajectory where we as humans will be involved in the work, but we'll most likely be working on another abstraction level. So, I mean, like so if you understand yourself as a coder today, it's not about in the future we will not need any coders, it's simply that the identity of the coder is not anymore, are you able to write code? No, no, no, that's the AI. The identity of the coder, can you develop great systems? Can you develop the right logic? Do you understand have you understood what is happening? So, what I'm saying is like my logic here is that the more AI is getting down and into and embedded in the workforce, doing the work that we as humans are doing, we are moving into other types of roles. We will have plenty to do as software engineers, but not the coding specifically. No, we will doing the other things that you're you know, what I was if you go 360 degrees around right writing a great software, one piece is coding, and then it's all the other bits and pieces to write the great product. And all this to me, I think I see a pat I see something that isn't this about critical thinking, judgment, reasoning that all stems from logics and all stems from mathematical thinking. So mathematical thinking brings becomes one of the par core, you know, system thinking, all this becomes core of what the humans continue doing, even if the systems are doing a lot of automation or augmentation around you.

Anders Enström 1:37:04

Yeah, but you asked earlier if what is the problem with math education today, and I think the biggest problem is that the student doesn't know what you talked about right now, so we need to teach them that AI can help you, and you need to be able to have some mathematical skills to be able to navigate in this crazy world outside. But it's it's the skols and the teachers and the principals and the municipalities that need to make sure that that is happening in each and every classroom, and how to do that? Yeah, that's a tricky question.

Henrik Göthberg 1:37:41

But are we shifting from calculus into reasoning? I mean, like another you know, problem solving, right? So, math, we can this we can understand math in different ways. Can you do algebraic? Yeah, yeah, yeah. Some kids can you if you give them an algebraic setup, they can solve, they can, they they can do it easily. Then you show them a problem written in text, and they have a very hard pack time unpacking what is the components of this problem and sorting in it. Yeah, both is math to me. Yeah, yeah, I of course, yeah. This is the part that without a script, without a problem, without a statement, you need to formulate the question. You need to be mathematically savvy to not to look at look at something very complex, formulate the problem, and then if you can formulate the problem and the reasoning around that, then you can feed that back into the AI. Sure. But to me, at the core here, reasoning is mathematical thinking or logic, or I don't know, problem solving, right?

Henrik Appert 1:38:35

Yeah, I mean, pro like I mean, it's complex problem solving, and mathematics is one domain where where we do complex problem solving. I think you you said earlier there are a lot of things about you know what is what does it mean to be a coder? And you know, I think and you talked about what's gonna what does that mean in the future, and my my question or my return question to that is well, when we say the future, are we talking are we talking 10 years or are we talking five years? Or exactly are we talking two weeks from now?

Henrik Göthberg 1:39:05

Yeah. I'm thinking slow moving. I I'm thinking like if we can see a trajectory, I think we can suck out slow-moving dimensions of topics that are super important. Uh simple logic, problem solving and complex sorting and problem solving will probably not go away as a problem. So this we need to be really good at, regardless of if it's five years or ten years.

Henrik Appert 1:39:29

I mean, so I mean, who knows how great, like to my understanding, AI is still not at the level of like really great, really complex problem solving. I agree. But but like I am also like I like just because it is not there now, like who knows where it is, uh, you know, and then if it's if it's six months or or or five years. Yeah. However, like I think I think that regardless if AI even does become better than us at uh than us at problem solving, it doesn't mean that we as human beings are gonna stop enjoying solving problems. Like, I mean, I'm sure there are listeners out there who enjoy playing chess or or solving suducos or crossword puzzles or whatever it is. And everyone who listens to this podcast knows that if they were to take a photo of that uh sudoku and paste it to Chat GPT, it would solve it like that. Yes, but we still enjoy but we still enjoy solving problems, like it's part of like uh what it means to be a human and enjoy enjoy our intellect and enjoy bouncing ideas off of one another. And you know, I mean I enjoy playing chess, and uh there are so many chess players out there that are a whole lot better than I am, and AI is definitely a whole lot better uh than I, but I still enjoy playing playing chess, and so I think that I think I think that problem solving is part of the human experience, and we enjoy it, and it should be celebrated.

Learning With AI Without Losing Thinking

Anders Enström 1:41:04

And I think that's that's appreciated. That's a valid point because I think that math education today needs to be more fun for the students. I think that's a crucial point in math education, it needs to be fun, and then you're gonna solve the sudoku or the cross-word puzzle because you like it. So I always have a task. I haven't had uh a math test in my classroom in 20 years, 21 uh and not any homework as well. So because I love you. Thank you. Not every teacher is gonna say that though, but I don't care. But I think that if you make math and science in this case fun and reliable to their own everyday learning, I think you're on the right path. And AI could help, AI could cheat for everyone, but you need to have AI as a partner and not as an answering machine. And I think if we understand that that it's not an answering machine, or if we're not treat it like an answer machine, I think that math education has a really bright future uh ahead.

Henrik Appert 1:42:13

But I mean I I I would be happy to go even further and say that this goes outside just math education. How do we think about like how we're interacting with AI?

Henrik Göthberg 1:42:23

Because one of the one I'm not trying to find this angle.

Henrik Appert 1:42:26

Yeah, sorry for interrupting. Yeah, but go, go, go. Because, you know, so one of the like AI can be a tremendous lever for efficiency and uh you know get new perspectives, bouncing ideas, yes, but but so here comes the here comes the really really important part. Like I think, and there's been studies showing the risks of this, that we can't stop using our own brains. No, no, no, and and and like just like well, I'm just gonna put this into my LLM of choice and it's gonna spit out the answer, and then I'm not gonna think critically about that. Do I agree with this? Do I think that this is does this make sense? Is it correct? Is it correct? Yeah, but and even if it is correct, even if it is better than what I would have done.

Anders Enström 1:43:13

Well, it's the benefits of learning in that.

Henrik Göthberg 1:43:26

Yes, yeah. This is this is I think we can this is my this is exactly what I'm doing, right? But but here now comes the interesting topic, then how we come back to how we make it fun and all that. And I had a topic here, rethinking, rethinking mathematical thinking. So, and the metaphor is a little bit like okay, the software engineering identity, I think software engineering will not go away. But anyone who put their whole identity as a software engineer into their I'm one of the most proficient coding coders in a certain language, and I'm brilliant at the code, you know, if that's not the identity of the software engineer anymore. And I'm thinking when we are now talking about rethinking mathematical thinking and how we talk about these topics, is about the cognitive abilities. So it's a little bit like with what you said. Are we needing to shift our way of mathematical thinking and what is important from uh oh pure calculus that you're really, really good at the the form uh the formality of math? And this is the math language. Yes, yes, that's that's that's the formal language, but math in the broader sense is problem solving, reasoning, what else can we talk about? Abstraction, yeah, all these things that are fundamentally so are we because maybe there is where we need to go with the mathematical thinking, and what is really important to stay human and stay cognitive, you know, to not lose cognition.

Henrik Appert 1:44:50

Yeah, so I mean, again, I'm kind of coming back to an earlier answer that I said, and I think for the broad masses, how we use mathematics, it doesn't change anything. I think AI changes how we can interact and how we can learn and experience mathematics. But when it comes to our day-to-day utilization of mathematics, how long time does it take for me to go from place A to B? How much money do I need to buy this? Like, just because we have AI doesn't mean that those things stop being a necessity. Like, so those things are like still going to be true, and then I think there will be those people at the at the tip of the spear of mathematical understanding and learning. We are still discovering and learning more things about the the universe of mathematics, and that is going to be enhanced and go faster with the AI, and there will be those spears in mathematics that are humans that can understand that, and and then it's going to be a pyramid. So there are like that single individual who understands this particular dimension of mathematics, yeah, uh, and that cascades down.

Henrik Göthberg 1:46:01

So I'm not sure if that answers your question, but it does, but it does it's not the answer I I think it is. Uh because in one way you say I think too broadly that it stays the same. I don't think it does. I think I think I think the argument you're saying is the fundamentals of math in terms of daily math that we need, that baseline is still there in used used to used to maneuver life. Okay. Right. Beyond time and in a meeting, right? Yeah. Then I think there is a new dimension of math that we are not putting enough attention on in because we are we are learning how to, I mean, like math in school to me is like the baseline, like this, and then a little bit more advanced calculus and and and and all that. And this is maybe important, but this is a little bit like for me. Yeah, you need to be more proficient in using a calculator, but you don't really need to understand this. But the other dimensions of math that I think sips in everywhere, abstraction, problem solving, and stuff like that, I think becomes more important. And and I think it's then how do I do abstraction problem solving in critical with the right of tools? Right. So it's so it's a little bit like so. If I go into the math, uh, you know, we're not allowed to use a calculator now. Okay, I go to university math. Of course, I'm using a calculator, a very advanced calculator, but I because I'm I'm focusing on the problem and reasoning and stuff like that, and I will see what you're having in university because because the tools are there to make the ground work, yeah. We need to bring the university math, which is this hardcore pyramid math. Some of that actually is needed to function on a middle manager level that is innovating their projects or their workflows for AEI, not to fuck it up. Yeah, so I mean again I'm I'm some you're right here, but I'm trying to pick out that mathematical thinking that might now be university level, that we now with tools, uh, we need to teach the kids problem solving. Sure.

Henrik Appert 1:48:01

I mean, I I guess one comment I have to that is I think I so the main reason most people train and and practice like multivariable calculus or linear algebra or you know, pick your uh pick your like strand of of university mathematics. It's not you know, it's very common among people doing engineering, and when they leave their universities and they go into the workforce, most of them go into problem-solving uh work, consultancy, uh banking, whatever it might be. Finance, uh finance and so on. But so, and so so I think again, like even at the university level, we are learning like we are using mathematics as a language to train our problem solving and our critical thinking and our ability to take something very complex and uh break it down into subcomponents that we that we can that we can then solve. And this kind of ties us back again to like to why it's so important to be showing your work to show your thinking. Exactly. Because in mathematics, you know, you get this like it can be a very hard problem or it can be an easier problem, but it's easier if we can break the problem into different parts. Yes, and it's so important that students get into the habit of of doing that uh very early.

Anders Enström 1:49:21

Yes, yeah, yeah, I totally agree. And I have an I have an uh like an assignment for my for my students each year that they they are gonna make their own math book, so they're gonna create problem-solving assignments. So we put it in a math book and then we hand it out to to year eight because we do it in year nine, so they can they can try my students' own assignments. That's pretty cool. If you can if you can create a problem-solving assignment, you also understand the problem-solving methods and how to do it.

Henrik Göthberg 1:49:53

So, I mean like so. We are we are we are moving into the final game here. Impact now and and even looking ahead a little bit, and we we we might be able to uh stop finish it off in not not not in two hours, but maybe two hours and fifteen. Okay. Sure. But uh because I'm I'm thinking now we we we are sort of coming back, which we have already had it on the table. I just want to sort of uh end here as a societal impact. We are talking about the wide uh of a divide, we're talking about a competence divide. And you can then also think about this potentially how can AI AI and ed tech either reduce that AI the divide in the in the world or how can it widen it? So do you see what I mean? Like it's like every like some argument some parts of the argument oh we don't want this tech. Um so I'm I'm trying to understand if you flip it, uh, the conversation on the divide, that in your opinion, the way I see it, you're describing this like we want to get these AI tools and technologies in, and we want it to make it for all to shrink in the to take away the divide. But there's an there's a counter-argument that uh technology is not helping or or sure.

Henrik Appert 1:51:10

And I think I mean so to me, the devil's in the details, and it comes back to how is technology being used, and if technology is being used in intentional, meaningful, research-backed ways, then you know, then we're gonna have a great, then we're gonna be in a great place. Yeah. And I think that there is a risk that because of the risk of using something in the wrong way, we say we're just gonna remove it, and then we don't get access to the tools, and then we don't have the potential like high uh the potential good outcome that we that we could be having. That's a short version of it.

Henrik Göthberg 1:51:48

What is your take? I mean, like if you look at AI widening the education divide or closing it. I mean, like for me, it's I I don't understand how it can widen it, but if you try to no no, neither do I.

Anders Enström 1:52:00

I think that I think that AI is trying to steel man the conversation a little bit.

Henrik Appert 1:52:05

Yeah, but I mean you can see how it like given that there will be those who have access and those who don't have access, then we are widening the AI to do it. Exactly. Sure. Exactly. So it's that we need to take it away for all, or how exactly. Well, so then we get into this landscape of well, how do we think about ourselves as a competitive nation versus other nations? And you can be like pretty sure that there are those countries that are going all in and thinking about well, how do we just reimagine how we have been doing things and how do we use this technology in the way that it's intended? I think Estonia is a great example of this.

Henrik Göthberg 1:52:41

But this is the way to think about AI widening the AI divide or not. It's literally very simple. We see a growing divide in the world in using AI or getting value from AI, blah blah blah. And this spills right down into our societies in terms of education. And you we will get to developing or developed countries in the same way.

Anders Enström 1:53:03

Yeah, but I think the way that politicians is talking now by removing digital tools from the classroom is gonna be that that the the child or the children that is from a kind of a poor home, they're not gonna get to use AI or digital tools, and the one who's from a wealthy home, they have all the access. So we are kind of dividing it just by taking it away from from schools, I think. So I think that AI used in uh in a good way is gonna democratize learning uh for each and every one in Sweden, and I think that's what we need right now. We want more engineers, but we don't want AI in school. So it's kind of a tricky question, I think.

Henrik Appert 1:53:48

Yeah, I mean you're saying that like both of those can't be true at the same time. If we want more engineers, if we want more people interested in technology, if we want more people leveraging technologies in meaningful and safe and intentional ways, then we can't say we're gonna give students who are over this age access and these students not access.

Anders Enström 1:54:11

Exactly.

Henrik Appert 1:54:12

So, but let's see, like I think uh there were also improvements to the AI commission's report because originally they did not mention um education, they cooped in education on one page.

Anders Enström 1:54:24

Yeah, yeah.

AGI, Purpose, And Human Agency

Henrik Göthberg 1:54:25

So that was positive. Yes, it was. So if we go even further now, as a as the last final question here, if we look into the future and we, you know, first of all, if we if we use philosophy, do we do we think we are on a trajectory that at some point we will reach AGI? Start there. I mean, like, so the distance That's an easy question. Well for me, I think I think it's an inevitable, but I can't I can't I can't put a number on it. But to think that it's not after the right, I mean like it's also about definition, of course. Exactly.

Henrik Appert 1:55:02

I mean, really again, like well, what do we mean by by AGI? But if we mean and you know, if you want to get philosophical and talk about consciousness, that that's a that's a different uh question, perhaps. But if we mean if we're talking about AGI as uh super intelligence, uh as some form of entity that is able to solve problems and do everything.

Henrik Göthberg 1:55:26

But but but AGI was the I mean like the same ultimate definition that we have circle in this part is a little bit like that the uh that uh the the AI can in in a meaningfully short time take pick up and do any task that any humans can do. Um so take over uh any job, which is really not at all worried about now, because we can have we can be way better in certain tasks or in or certain things that the AI is doing way better than us, but to really do everything that Henrik is doing, it's a long way away. But but that's the first level. Then we can talk about super intelligence, uh something else. But but for me to reach that, you know, what the average person can do and can manage as a coworker, I think we will reach there at some point. Um for me, that's the first question. Yeah, so it's the question, do I think that we are going to get to that point? I just want to do a calibration, do we believe in it or not? Because my question is based on that belief.

Henrik Appert 1:56:30

I mean, it's it's I think it's pretty inevitable. And then the question is, is it um is it is it three years away? Is it five years away? Is it ten years away?

Henrik Göthberg 1:56:41

So let's not go into that one, but with that baseline that it's inevit inevitable, what role do we think human learning and education will have when that's a fact?

Henrik Appert 1:56:51

So the same role that I alluded to earlier, like I still play chess, even though computers are better at chess uh than I am, and I still enjoy learning and I still enjoy thinking, and uh I think me still being able to you know use my own agency and make decisions which I'm going to do better if I have a better logical thinking, then I think that that is still important. Like I don't envision a world where I have an AI who just tells me to like now you're gonna go to the now you're gonna go to the bathroom because you need to pee and you need to leave right now in order to catch that bus. Like, I think people are still going to have agency.

Henrik Göthberg 1:57:34

And um are you are you on this same track? I'm I'm very much on your track here.

Anders Enström 1:57:38

I am, and I think that and I think the human conversation is gonna be much more interesting and much more fun and much more important. Um like this sit down and talk about AI.

Henrik Göthberg 1:57:51

So I think that's gonna be we we know the answer when we look at Star Trek, right? Yeah, they're like they still go to their Star Trek Academy, right? Exactly. Yeah, yeah. So that's the point is this even when we all this magic can be done, the sacredness of the classroom I think will exist. What do you think?

Henrik Appert 1:58:11

I mean, I am certain that why not? I mean, again, I I come back to us being social beings, we enjoy spending time together, we enjoy solving problems together. We um even though a computer could do something faster or better or in a more efficient way, people are still going to want to do that. I think with you know AGI, there will be other societal challenges uh that come. Well, how do we set up a society that is structured in a way to support this transformation? And you know, the the closest, the closest thing to my knowledge that we've had to this type of like societal complete uh disruption has really been you know in industrialization, you know, in the in the in the 1800s where people used to be living on farms and now suddenly everyone moves into the cities, and you know, you have spinning jenny and all of these innovations, and it was uh a complete transformation of like what everyone's lives looks like.

Henrik Göthberg 1:59:13

And I think so. I've I've used those macro cycles agriculture to industrialization, and industrialization slowly transitioning into some sort of information age that hasn't really been a paradigm shift. But I think you know, if if you think about it from agriculture to industry, and from industry to AI for society, I I think we will I think that the impact is equally big or a big thing.

Henrik Appert 1:59:33

Yeah, but it's but but I think one thing that is you know like it's like what's what's what's what's easy about the first analogy about you know from from farming to industrialization, like people are physically moving from one place and they start going to factories, and our rather simple brains can like we can make sense of that. But what happens when a computer can do the job that you are doing more efficiently and create value more efficiently? It's just like I think it's it becomes an uh harder to imagine, right? I mean, I think it be like I think we human beings, and now we kind of get into the philosophical in the philosophical space now. But in the philosophical space, we're like, but I think human beings need uh need to feel that they have a meaning.

Henrik Göthberg 2:00:28

Yeah, a purpose, a purpose. Exactly.

Henrik Appert 2:00:30

Like and I think Will that go away? Of course not. I don't think so. That is Well the need won't go away. The need the need The need to feel purpose. The need for purpose will definitely not go away because again, like we are the same, we are the same apes that we were when we you know climbed down from the trees uh a couple of uh hundreds of thousands years ago, right? Uh so like that need is still going to be there, but our ability to to feel and derive a need uh or or feeling purpose and need from like having a work, like well, who knows what that looks like. And um, you know, it could be glorious and it could also be pretty uh cataclysmic and and scary and uh not so glorious, not so glorious, and like the verdict is and now we're not even getting into the space of what happens if AI takes over and becomes our overlords, and you know what happens to the like will will we ultimately survive? Like now we're not even talking about that. Um but that space. And I don't think that that's a like an like a zero probability that that happens, and that is a scary part, but no, and so we had that.

Keeping Curiosity Alive In School

Henrik Göthberg 2:01:36

That's actually let's park that question, but because that's that's the traditional final question where do you put yourself on this scale on the spectrum from utopia to catak, you know, to catastrophe? Yeah, because there is the you know, where but let's that's the final question. The one question in between will mathematics remain central to how humans understand the world with AI fundamentally changing everything. So so will AI change our fundamental relationship to knowledge and and problem solving and learning? If you look at if we look at it at the broader sense, math is at the cornerstone of understanding the physics and the biology and everything like that. Right.

Henrik Appert 2:02:19

Well, so the math isn't changing, right?

Henrik Göthberg 2:02:21

You know, one plus one is still is still two, and you know, no, but uh but now we we we rely on it and it's deep in what we do, but if you imagine the AI is doing it way better than us. Well, so So how does that change our outlook on these topics?

Henrik Appert 2:02:36

Well, so what I think is important for the education system is that we keep on advocating for the importance of driving understanding and curiosity because I I think that a society where people are informed and are using their brains and their critical thinking and their creativity, and these are all these are all skills that we train and develop with mathematics. So, but do I do I fear that there will be portions of society and where math knowledge goes down and their understanding of the world goes down? Do I think that there's a probability of that? 100%.

Henrik Göthberg 2:03:15

I think they're not consciously working against that risk.

Henrik Appert 2:03:19

Like we will end up there, and for some like people, they will end up there, and it's a tragedy, it's a tragedy for themselves, and that if it's if it's enough human beings, it's a tragedy for society as well. What's your view here on this?

Henrik Göthberg 2:03:32

I mean, like, so I think you're spot on.

Anders Enström 2:03:34

Yeah, I think so too. Because, yeah. Oh, that's a big question.

Henrik Göthberg 2:03:39

That's a huge question. That's the fun part with the last questions. They're super big.

Anders Enström 2:03:43

Yeah, it is so, but I think that that math in its core is gonna be super important, uh, because you need math to navigate in this crazy world and in your daily life. So I think that's not gonna just disappear. And math has so many life skills.

Henrik Appert 2:04:03

I mean, to me to me, math is just the word that we use to describe the subject of logic, logic and problem solving.

Henrik Göthberg 2:04:13

Uh, everything I do as a as an entrepreneur is problem-solving, abstraction, and sorting on some level, right? Yeah, understanding my priorities, understanding my constraints, understanding what I think. That doesn't mean that everything is mathematical.

Henrik Appert 2:04:28

There's a lot of emotional intelligence and collaboration and things that get into entrepreneurship.

Henrik Göthberg 2:04:34

But the way I make sense of where I should go, I I don't I'm I'm not math, I'm not mathematical at all, but I realize I'm sorting, I'm I'm abstracting, I'm generalizing, I'm using so you're doing mathematics, I'm getting mathematics. You're using probabilities and you're using and and maybe we don't even need to know the technicalities of it, but we need to be taught of thinking, yeah. I don't know. So I because I'm I'm literally scared that when you're relying too much on the AI in the wrong way, your thinking is not as sharp. Or you look that if if you're not actively working on this on conscious, I mean it's like if you're not conscious on what what the hell is that?

Henrik Appert 2:05:18

If you're actively not if you're not if you're not conceptually thinking about what it is that you're using, if you are just if you are just consuming AI without engaging your frontal cortex.

Henrik Göthberg 2:05:31

So this is the difference between using Chat GPT in the right way or the wrong way. Yes. I can get so much value from Chat GPT to do something that I is originally mine 100%. Yeah. But but you know what? A lot of times there is a I start with the one sentence and then I need to tell them to shut up. I don't want your answer on this bullshit. I want you to help me fine-tune my idea. Yeah, this is the difference that some people, if you if you consistently use take that first thingy. I mean, like if I'm super lazy, I I said, Oh, I need to I need to prepare, I need to prepare an AI podcast with Um, you know. He started off without me even asking him, right? No, I want to go in this direction. This is the danger I think that AI. Yes. And this is where education needs to sort of keep us curious or keep us conscious of what is the real thing you're supposed to do.

Henrik Appert 2:06:26

So, how do we how do we develop both identity and agency among our learners? Like, how do we how do we develop people to be curious about their own thinking? And here we get into spaces of like mindfulness and you know consciousness and being relevant, general well, like all of these topics.

Closing Thoughts And Cliffhanger

Henrik Göthberg 2:06:46

No, but it's about also like you can go into different areas here, like knowing yourself and you know, understanding what you are doing. I mean, like it's like looking at yourself from the outside, understanding uh uh this is deep. What do we have answers to this, or this is another two hours? I think I think maybe that's a different conversation. This is the topic for the AI after work when we turn off the cameras, drink more beer, and solve the real problems. Great. Should we end there? Yeah, we end on a cliffhanger. Yeah, yeah.

Henrik Appert 2:07:20

And Shannik, this has been a pleasure. Really fun. Thank you. Thanks. Thank you so much.

Henrik Göthberg 2:07:26

Likewise.