Episode 13: On the Near Future of AI

[MUSIC PLAYING] You're listening to Community Crafted Education, where we explore alternatives to conventional K-12 education and the potential impacts on society and culture. I'm Eric Wairn, director of the National Hybrid Schools Project. On this episode, we're joined by Matthew Sanders from LongBeard.com. We'll talk about the present and future of artificial intelligence, how it could go right, and how it could go wrong. Matthew, thank you for joining us today. Thanks for having me. Yeah, so maybe we'll just start with, why don't you tell us a little bit about yourself and LongBeard? Sure. So LongBeard-- let me start with LongBeard-- it's a company. Essentially, we build and scale Catholic guy. That's essentially our mission. We kind of got started in technology more generally about 10 years ago. We initially were just kind of building websites and web platforms for organizations around the world, namely Catholic organizations. And that eventually led us to Rome and working for the Holy See and how the universities and things look better on the world. And eventually, that led us to some interesting projects. One was working for the Holy Father and studying at a technological infrastructure for a section that he ran. And that got us in collaborating with Google, which got us interested in kind of emerging technology. And basically when we saw Chattanova T, when it dropped, we found out that Catholics were using it. That's kind of when we decided we wanted to see if we could find a way to use the technology to serve the mission of the church. Yeah, great. So we host here a big conference every year when we bring together all of these various hybrid schools and micro schools and people interested in these things. And one topic that emerged over and over again this past year was AI. So that's kind of why I was interested in talking to you. And so we'll start as far as questions go with this. One reason that a lot of these schools start is because the founders who are often parents will see or maybe hear something that they don't like in conventional schools. And so they'll start a school to address that. Or maybe from a more positive direction, they have some idea about the good society that they want to promote through a school. And so they just start one up and open it. So you're working on AI that can sort of faithfully answer questions about or from the perspective of the Catholic Church. Another company recently announced that it was going to allow porn creation on its services. So I don't think it's surprising that some school leaders can be sometimes skeptical about letting AI kind of influence their students. I've seen other tests where they'll come out and they'll find that various AI's have different kind of default political leanings. And that might be fine for some schools. And other schools might not like it. So can you talk a little bit about this about how we decide or manage things like AI's truthfulness or hallucinations or overall worldview? Sure. I mean, it's certainly a big challenge. I think the first thing to acknowledge that not all the AI's are made the same. So it's interesting because a lot of the companies are using very similar pools of data. But there's a bunch of what we call post training they do on the model. And that kind of shapes the personality and kind of the guardrails of the model. And that influences its overall behavior. So when evaluating which AI to use, there are benchmarks you can look at, which evaluate them on certain areas, the ability to effectively answer questions around sciences. And general knowledge questions. So that's one way if you're looking for technical benchmarks to understand their capabilities. But of course, the other way is just the vibe test. You just have to try it yourself. And I usually recommend that if you have a particular world, if you're your particular values, or let's say you're a Catholic, and that's very important to you, you can come up with a list of questions yourself. We call these evals. And you can just ask the model on issues that are important. That you think there's possibility that may be kind of divergent. And see what it says. See what it says. Sometimes it may refuse to ask. Sometimes it may not. So in every model, depending on, you know, that the company behind it, and what their particular kind of values or guardrails are. You know, I mean, in the early days, when we got into this, this at least happened a lot. Over time, I think the models have gotten better, but there's still some innate biases in them. The last thing I would just mention is, it's important to understand the companies behind them. I think this is really important. So why was this company founded? Who's leading it? What's their track record? When it comes to political biases, religious biases, and things like that? It's just a company that, if somebody asks about AI, I was using it, I'd be proud to see. I'm using them. And you know, if the answer is no, then you may want to think about using another company. Yeah. And well, there are still humans involved in this too, right? I'm going to get over my skis very quickly here. So stop me when I start to, but it's not yet the case that the AI's are sort of independently, entirely updating themselves and evolving on their own. Yet, there's still humans at some point either choosing the data that it's being trained on on the front end or, and/or, there are sort of human testers on the back end telling the AI its answers are good or bad or right or wrong, right? So there's still some human presence that you can also take into account a little bit, right? Yeah, I mean, absolutely. I mean, I mean, I mean, I mean, you hit the nail in the head, Eric. Ultimately, the models are trained on human data. And so their biases will reflect the biases that are represented inside of the data. And then in post-training, there's a bunch of different approaches to post-training. And one is, you know, it's called supervised fine tuning and there's direct preference optimization where the human being is well, basically, the AI will answer and it'll tell them which one it likes best. And so there's also reinforcement learning, but ultimately, it's trying to steer the model to answer in a particular way. And as you say, humans are more involved and like the direct preference optimization is DPO approach and they are the reinforcement learning, which is more about, you know, providing basic verifiable rewards to the model so it can kind of learn on its own a little bit more independently. But there's always a human in the loop. Yeah, great. So these schools as they're kind of deciding how much or in what ways they want to interact with all these new AI products that are coming out and being baked into maybe products that they're already using. These community crafted hybrid schools and micro schools are an interesting case because they can be really nimble and at the same time, they're really high identity, right? So they're, we're not talking about adoption for a big, you know, a 10,000 student school district or something like that. We're talking about much smaller scale. The fact that our high identity means we're also not always trying to be the most moderate or cater to every single opinion, right? They can decide which way they want to go. They also don't have to care so much about particulars of the state curricula or maybe not at all about state tests depending on how they're set up. So these schools have a lot of area to maneuver, right? So I see these schools shaking out in the kind of three groups on AI. So one group is all in. So they wanted, they want some sort of AI to be kind of directing their curriculum day to day, month to month. They want the students interacting with it all the time. That's gonna affect how their schools are set up. There's some schools in Texas that are realizing we can, because the AI's are so well able to see where the kids are. They can really compact their school days into much shorter days and give them time for other things. And then there's a second group that's very skeptical, right? So they're saying, no, no, we're gonna double down on paper and pencil assessments and activities and assignments. We're gonna ban it outright, although maybe we can talk about why that may not even really be a possibility for the most part. But they're kind of going into this slow, deep learning approach as kind of a counter to AI, right? And then there's a third group that's just gonna wait and see how things shake out. So I wanna focus on the first two extremes. How, if someone came to you and said, Matthew, I want to start a micro school and we're all in on AI. What kinds of things might you say to that founder to advise them as they're getting going? - Well, I mean, I need to have to ask them what they mean by that. I mean, you know, effectively you have to really get down to what is the T-lilson education. So what are we trying to do here? I mean, in principle, we're trying to ultimately help some young human being to flourish, right? And part of that is making sure that they have solid human intellectual spiritual formation. And obviously, if there's AI, you know, maybe we'll take a box, but the other two boxes, I don't think effectively. It'll ever really be able to take, right? So, you know, I think if it's about, I'm parting knowledge. I think the AI is gonna be very good at this. So, you know, as these AI's get smarter, they'll be able to adapt their pedagogical approach for each student individually. This is why platforms like Khan Academy, which is, you know, incredible platform trying to serve people and develop them, which is difficult to access high quality teachers are using AI heavily because it can patiently work with the student and ensure that they're understanding, right? It can move at the speed of the student. So I think there's a lot of possibilities in that area. But at the same time, you know, it's more to school than just parting knowledge. And I think in some ways this actually may free us from the little educational drift that's been happening where the classroom is that I can be converted from a place. This will be that we're just kind of facilitated learning going on, right? Where the students are very interactive in the educational process. They're learning from each other. To a place where basically one teacher just kind of does a download for the students and everyone's kind of expected to just learn it. And so I think in this regard, there's a possibility for a good balance here, right? Where students may have a private tutor for learning. And then they have the opportunity to be in the classroom so they can learn from others. Right? They can, they can, there can be a problem. They all share their unique perspectives on how to go about solving that problem and the teachers basically helping to facilitate. Provoking that whole process. And I think that might actually help us get back to a form of education ultimately might might be more potent and effective. Yeah, I can also see how this. At some point might enable some schools to come into existence, right? If you can't find a particular teacher. To teach a certain subject, maybe maybe the AI will be good enough to kind of fill in for that on a technology role. And then you have a facilitator who's a human on the other side. Now, what about the skeptical schools, like someone comes to you and says, look, you're, you're a great guy Matthew. We're going to open this micro school though and it's going to be, you know, paper and pencil and oral learning 100% of the time. What, how would you, how would you talk to that kind of person? Well, listen, in some ways I, I feel the draw. You know of that. I can, I can understand the appeal. But in the end of the day, school schools will be about, and this is meant to be, you know, a complimentary process with the parents is to make people discerning who they are. And what they're particular gifts are and honing those gifts, right? So they can basically use them to, you know, to serve others and glorify God, at least that's how we look at the Catholic perspective. When you start making some certain presumptions, right, or you start closing certain doors for certain students. I don't think that's necessarily a good thing. I mean, I really think that education in the future, because I think we'll have these, these AI tutors and things like that. We should, we try as much as possible to facilitate broad spectrum exposure to as many areas as possible. So essentially the students have an opportunity to click with something in particular, but they feel like, wow, this is it. This is what I was meant to do. And I think it's first people like, you know, myself, but I wouldn't say I really clicked. I mean, it took me, it wasn't slow. I think it was, you know, middle aged by to my phone. I found something. So, so I, you know, I think if you, if you're basically coming at it and saying, yeah, AI is just not something we want our kids exposed to because, you know, we have this kind of clad more classical understanding of what education is supposed to be about. And that means paper and pen and things like that. I think it's some ways you're maybe hobbling certain students. So I think that there's a, there's a happy medium here. I like, I'm not a fan of devices being used in the classroom setting. I think that defeats the purpose. So I am, I could see why while they're in class, you want to put away their phones. And you don't necessarily want an AI necessarily dynamically, this is being a responding to the classroom. And sometimes you need to ensure that they're exposing the students to it. So if they actually, if there's something about AI that they're drawn to, they have the opportunity to explore that and and be nurtured. Yeah, oh, that's great. So you, okay, so this leads me to another question that I have for you. You mentioned the idea of closing doors. Right. So we'll, we'll talk in a minute about where all this is headed for for the bigger picture. But assume, you know, in more and more powerful, AI is coming. Right. So what kind of messaging should we be working on for students. So for example, it's take this for what it's worth since I'm asking it as a university professor, right. There's been this argument that not everyone has to go to college, right. So there's some, there's some merit to that, I think, but it can be kind of a dangerous message, right. If you tell a 13 year old boy not to care about college, the danger is he might listen to you. Right. So that can be okay or more than okay for many of those kids are even most of them. But there is a danger that that kid will listen to you and mentally check out at 13 years old and then say he graduates at 18. And I mean, it's not uncommon, especially for guys to then wake up one day and they're 28 years old and like nothing has happened. Right. So if you're saying we're going to these are going to keep improving. What I hear all the time is we're going to have all this idle time to focus on other things that we want to focus on with our lives. How do we prepare students for that. And maybe as a species, are we even capable of handling that kind of thing. Yeah, I mean, it's a tough one. You know, I generally think with the right formation. So, you know, if a person, you know, from the youngest of ages is well formed. And again, looking at the three buckets of human intellectual spiritual formation. Then I'm not worried. I think ultimately provided that they have an environment in which they can they can truly explore the different ways that they can add value. And again, that should try to make that as broad spectrum as possible. I just generally think that they'll probably find something that they feel like they're meant to do. And that will be a transit field that will be motivated to do it, right. It won't be it's not really going to be a because of money. Or because someone told them to do it. They just it comes naturally to them and they just feel a great sense of peace in doing it. And I think that, you know, I think that's that's that's very exciting, right. I think having having the time and recognizing people, you know, they may take more while to figure that out, but they have the time there isn't pressure for them just to kind of level up. So, you know, so they they can then, you know, go to university, pick a major. I think that that's that's a good day. I don't think, you know, telling anyone that they should go to university or not go to universities a good idea. I mean, I think people have to discover this on their own. I think we simply, I think the purpose of education is to put possibilities in front of people and then allow the opportunities to explore those possibilities and hope this hope that eventually they'll find one that you know, just kind of clicks. And if we don't immediately, then we it's a question of being patient and encouraging to keep going, right. So I think if we look at in that regard, I think we should be fine. But again, good formation is not easy to deliver. It's exceptionally difficult. I mean, I spent years in seminaries, starting to be a priest right. And you know, it takes five years to basically get a priest ready to be ordained. And that's intensive, right. I mean, you're looking at comprehensive components of the individual human is lost. And I think about how many of the average lay people get that kind of intense formation, right. Preparing them for their label case, not not many. So I think we have a lot of work to, you know, to do in this. I, you know, I do generally think with AI and robotics that you're going to see a lot. I mean, most areas of the GDP economy will be fully automated. So I just I don't see. I mean, if, you know, if you're, if you're thinking, you know, I have this idea of what I want to do with my life and it involves plugging into the GP economy and, you know, changing the world. I mean, I do think you have to be very circumspective about that because the world's changing so quickly. And that's why I think it's more important to just know what your gifts are. And, and, and be it be a, be it contempt just just use them just because that's what you're meant to do as opposed to focusing too much on a particular outcome you want. And that could be make a change in the world or it could be making money. And I think both of those are going to be difficult. I think in the an robotics age. Yeah, there have been multiple articles that cannot recently about how the computer science and engineering majors are struggling much more than we maybe would have expected them to five or six years ago. So maybe the field is broader for you to look at. Okay, so you're, so you're definitely pretty, maybe hopeful, but that's not the word I'm looking for maybe bullish on the prospects for AI and society over the next five or 10 years. So what's your, what's your argument that things are going to go well. And we won't get either sky net from the terminator where we have violence or a wally where we have super abundance, but everyone's just kind of slothful and without a purpose. Well, I mean, I'm bullish in the sense that I do believe that the Catholic faith has has the answers, you know, to, to how do we, how does humanity going to make it in the antibiotic stage. So I'm bullish in the sense that we don't we need to do here. I'm a little bit more, I don't need the word cynical. I think all this is pragmatic. Is that imparting delivering good formation at scale. We've really struggled with that. And I think that we're going to effectively have to get a lot better or there's a real risk we're going to have high casualties. We basically, usually when I talk about this, they kind of, there's kind of two ways we could break this kind of the golden, the golden path and, and there's kind of the dark path and a golden path would be, you know, we, we recognize we have a very honest conversation now about these technologies where they're at now and where they're likely to be in five and 10 years. And with that, then triggers a very robust civil conversation. So we all become aware of where these, these tools are at. And then we drive a conversation, you know, where's where these, where these development is going to take us. How do we feel about that? Do we want to maybe, you know, talk about specifically about what kind of vision we want for for our societies and one for for our kids. And so basically once we have kind of a clear sense for kind of that at the end state, we can kind of work backwards and say, OK, so what do we need to do what's techno what we need to point these technologies in order to arrive at that better world. I think it for effectively able to do that. Then I think I think we'll be fine. And I think this could this literally could be a golden age for humanity. If we don't, so let's say that we don't have that robust civil conversation, we leave the managing of this transition to the antibiotic stage to politicians, I suppose, and corporations. I think there's this real risks there and large part because I, I don't think those companies or, you know, I think most politicians are going to be equipped. that the change would happen to you. fast and I think especially for our elected leaders, I don't think they really have a good centrist going on and basically the companies are kind of driving just transition and although they may be led by people who believe ultimately these technologies will do great good, they obviously don't have a two comprehensive understanding of the overall impacts. And so there is a robust conversation meaning that so the civil society isn't really engaged in this process and effectively as greater and greater automation comes into our world and more than people are basically out of the job. We don't know how to help those people because we don't know how to impart good formation and so they're left with budget time on their hands and at this sense of they don't know the uselessness, right? They don't know who they are, they don't know what their gifts are and maybe we find some way to keep the pitchforks out of streets by giving some UBI, right? So they're not losing their house but they're just sitting on their couch and this becomes an extensual risk, right? Because then people may say, well, I'm just going to go with the gaining or an eminence chase other dopamine, you know, pornography and everything else and that just drives civilization basically right out of a cliff. So I do think that we're in a very critical juncture and this is one of the reasons why I think it's important that everyone has, as at least engaged, is looking into researching AI and robotics and again, if you're not finding good answers, the demanding good answers because I think the last thing we want to do is should be passive. This is not the time to be passive. I think every one of us has a responsibility morally, I think, spiritually to be engaged in what's happening around us. Yeah. So does your goal of path then also include some kind of universal basic income but maybe more like eyes open about it ahead of ahead of the time and decisions about formation and what we should be doing with our time since we can't, you know, making money like you say is going to be much more difficult, but you're not going to need it as much. Yeah, I mean, I, it's difficult to imagine how we can make this transition without some degree of, of UBI. Now, that may not be necessarily income, that could be resources and kind, but I do think that the robots within a few years, they'll start manufacturing those at scale and blue color will wake up and see that, you know, all that, all that factor working out of anything else, robots are coming for those. And I, as much as I think it'd be nice to be just pause and take the time to kind of work through this, you know, a little more deliberately. I think the economic forces are just going to force these companies to automate in order to compete. And so I think we'll have very little control over this process. So I think we need to make sure people don't lose their homes, that the parents know that the kids are going to be fed and that they still have hope. And they'll be able to be who they want to be. And I think if the government can find a way to do that, that buys us a bit of time. But again, that will only last, you know, it's only -- it's only -- it's -- it's only -- it's -- it's only -- it's satisfactorily. It's a minimum extent. Right? Because ultimately, we need to be usefully constructed with our time. And this is where I think the government on its own is -- it's not sufficiently in part that -- that formation. I mean, even in green and what that curriculum she even consists of, I think, could be -- could be disagreements on that for years. So I think organizations like the Catholic Church, if -- if -- you know, or any organization, if they -- if they -- if they feel very convicted, that they have answers to these -- these big questions that people are not asking, that they just have to work really, really hard to try and reach as many people as possible. I mean, I think the Church's evangelization mission, obviously, has always been -- has always been critically important, but it's -- it's -- it's -- it's -- it's -- it's going to be very -- it's going to really be going to focus on how important it is now, because if -- if we don't effectively succeed in this mission of evangelization, the casualty rate will be much higher. I don't think we'll be able to escape the fact that, you know, I just didn't -- I just didn't reject that the one person and -- and talk to them and share. And now that person is -- is decided they're going to offer you the nation. Right? I mean, I think -- these are the kind of scenarios we're going to see more and more of. And so I think that -- I think it's just -- it's important for us to take -- take stock of that. >> Yeah. The purpose is going to be much more important. Your purpose is not going to be the rat race in the future, because there won't -- there won't quite be one. Is what -- is what it sounds like. >> Yeah. I mean, I mean, there's still -- you know, I often say, you know, I think 80% of jobs in a -- you know, I'm talking about like attending a horizon will probably eventually be automated. Maybe a little bit longer than that. But there still be 20%, but I think that 20% -- you know, it's going to be highly, highly competitive. And given -- given some of the trajectories, these trans-humanist trajectories where people were talking about, like, merging man and machine, like neural link, you know, saying my computational capabilities -- I mean, so I can compete for this, you know, 20% of these jobs. I just think for a lot of people, it's just -- it is not going to want to become that in order to contribute to the GDP economy. And so that's why I think we have to provide viable alternatives. We often jokingly call, you know, the Etsy economy will merge. Well, people will just choose to pay more for goods and services because they're made by human beings, right? And people are contributing to the Etsy economy, again, not because they need money, just because they just above creating the things, ensuring them of others. Right. Right. So you've handed it this a little bit, but, you know, I think you're kind of channeling this Peter Tiel comment that we've made a lot of progress in the world of bits, but not so much in the world of atoms, right? So we've got all these video games, social media and apps, but that's all kind of vapor. So the question is, where are the flying cars? Where are the non-murderous robots? Right. So it sounds like you think those are definitely coming. So what -- how do we deal with that? How might that affect schools and kids themselves? Well, I think this is really going to come down to sales. We have to be able to sell, you know, the Catholic way of life. Like, we need to -- hey, listen, if you put this the time and energy into this, it's going to be hard. It's going to be an adventure, right? You're going to stumble along the way. But if you put that time and energy in and you persevere, this is what your life can look like. And look at that. Doesn't that look great? Right? Because if we don't effectively sell that, you know, what the world's going to be selling is gaming and things like that. Where, hey, I get to invest time and energy and, okay, maybe I'm not necessarily upgraded, but my avatar is, and it's still kind of social and it's fun and I feel like I'm progressing, but we know ultimately where that's going to lead. Right. It's not going to be enough. And so this is going to be very difficult, right? I mean, gaming offers immediate gratification. It's fun, right? I mean, it's -- so -- And this is where I think, like, we have to do a good job of -- like, you know, I have to joke, like, back in the days when the lead of the Viva was around, right? I mean, the kind of the nuclear family and what life was all about was kind of brought in to kind of pretty clear focus for a lot of people. And so, you know, obviously it looked a little bit different for some, but -- but ultimately we kind of knew what life was about, right? It's certainly a vocation, you know, if you're a lay person, it's finding -- finding your spouse having those kids, investing and ensuring that those kids, you know, grow up, right? And living life modestly and putting -- putting Christ in the center. But there's a generation or two that have grown up without that model. I mean, I mean, it's difficult for them to even imagine what that -- with that Catholic Christian way of life looks like. And that's a real problem. You know, Colonel Collins, my old boss, he's to say, if you know where you're going, it'll be more likely to get there. And I think that really kind of sums up the problem and the challenge. Yeah. For everybody, right? Just knowing who you are and where you're going is kind of be even more important in the next -- I mean, we talk about five or ten years, but it's time to get on that right now is what it sounds like. Yeah, that's my -- I think we have to shut that from the rooftops. I mean, five ten years, and some ways will be too late. And these trajectories are very difficult to move off of. And the decisions about what America is going to look like in five years already be made today. And this is why it's so important that we -- you know, that we let our voices be heard, and we basically kind of raise our IQs on these issues. Yeah. And this is why I think it's important for all of these various tiny schools around the country to think about this, because as you say, politicians are likely to be behind, but your kind of small local communities maybe have a better pulse on where you want your own, you know, your kids and your families and your friends to go. Yeah. And I think those smaller schools, I think, you know, in some ways, they're ahead of the curve. I think a large part was probably going to happen is cities will start to empty because we won't need them as much, which is probably a good thing because they're -- you know, cities are -- they're a post-industrial revolution creations, right? And they're inherently alienating and isolating. And, you know, they're moving from the earth from nature. And it's just probably not good for us, net net. I'm not saying some people, you know, they may thrive there, but most of us, we could probably do it without it. And so, and I think the part of the thing that government's going to have to do is to kind of transition people into the world of heartland again. And the question is when people are moving out of the cities, you want as much as possible to be very intentional about forming communities, right? You want to kind of agree, what is the organizing principle that we're kind of uniting around? You know, if you go to Europe and you travel on medieval towns, it's pretty clear what the organizing principle is. It's a church of the center, what builds around it, right, or a monastery. And I think this is going to be, you know, very important in the future. And so in some ways, you know, the kind of the homeschooling communities, that's essentially what they've done. And they decided that they want to take ownership of the formation of their children. And they're going to decide that, you know, if they're Christian ones, this is crisis at the center of all of this. And so I think in some ways, like, they're -- they're the ones who are probably -- they're ready to succeed. We just need to make sure that what they're already doing, the benefits of that are shared very broadly. And there's kind of a toolkit that the others can adopt and use. Yeah. Great. Okay. So last question then, Matthew. What's the most interesting thing that you're working on right now that people interested in all of these hybrid school and micro school startups would like to hear about? Well, I mean, you know, I probably were probably most known for us called magisteria and AI, right? And it's effective. What we're trying to do is digitize everything that churches ever said. So right now, I think we have over 29,000 works going into these points of the University of Libraries and digitizing them. And then basically taking all that knowledge and making sure that anyone who wants access to it and get access to it. So kind of like, you know, a magazine that feels like a church meeting product, but what's unique about it is that everything that you ask, the NAI is retrieving source documents. And then. essentially making sure that those also formed the grounded answer. And so what gets me excited is that, you know, never before in history, the history of the church, have we been able to kind of distill from everything that you've written insights for a particular individual, right? So you have a burning question, and now you can ask it. And within seconds, you basically can have a consult with Thomas Aquinas and Augustine, the falls of the church. I mean, this is pretty wild, right? So if human intellectual spiritual information is critical in order to ensure that humanity makes this transition effectively, I think having tools like this, where we can kind of grow and learn from some of the greatest minds have ever existed. I think this is gonna be a real competitive advance in trust. - Yeah, okay, great. Matthew Sanders, longbeard.com. Thank you so much for joining us today. - Thanks for having me. - Thank you for listening to Community Crafted Education, a podcast hosted by the National Hybrid Schools Project, and Kennesaw State University's Michael J. Cole's College of Business. If you have any questions about the topics we discussed on today's show, or wanna learn more about what we do, please look up the National Hybrid Schools Project or reach out to us directly at coalscollege.com/hybrid. (upbeat music)