Hello

And welcome to the Moodle Podcast.

Hi everyone. I'm Abby Fry , and it's nice to be back on the Moodle podcast. My guest today is Moodle's founder Martin Domas . Firstly, a big welcome to you, Martin.

Hey, Abby. Uh , always a pleasure to sit down and have a chat with you and , uh, with the wider community. And , uh, so it's good to be back on the podcast.

Yes, great to have you. And I thought today we might start with just talking a little bit about where you are. I think it's really nice to place someone in their environment. So why don't you tell us where you like to work from?

So, I work from home as I have done for a long time. Uh , I have a , uh, a new house I got about a year ago, and I, I'm loving it. I feel like I'm gonna be here for a very long time. So I'm looking out the window at the Australian bush , uh, full of kangaroos and , uh, all kinds of critters. I'm , um, pretty well set up here to just work. And the way I'm working now is much less meetings and a lot less of things I can do in my own time . So my , uh, I'm definitely not working a nine to five . I , uh, my, my natural inclination is to go into the late wee hours. So I've been working kind of seven days a week, really around the clock in patches here and there. And , uh, that's really good as I sort of chase the inspiration. And

I was going to ask you about balance, actually, because I know that you are inspired by nature, or as you said, you like to be around nature with the bush. Yeah . Um, and being out of the , um, heart of the city. Do you have balance there , Martin?

Yeah, very much so. Yeah. Um , um, I'm really balancing things quite well. I feel, and I do use ecosystem language a lot around Moodle. Open source in general is a way of binding an ecosystem together, and I think there's a lot of inspiration from nature about how we should build our technology and how, how we should build , uh, companies and organizations and things as well. So , uh, yeah. Yeah, it all, it all comes together.

That's a nice segue into our conversation today. Of course, your Moodle's founder, and you've led Moodle as CEO for the last 22 years or so. Last year, you made a decision to transition from that role to the head of research. What motivated that move?

Quite a few CEOs seem to be doing it at the moment, or maybe I'm just noticing them, but how it was for , for me personally, was just excitement about these new technologies that are finally coming together. And I've been following AI and augmented reality and things around there for the past 20 years , uh, at least. And only recently have they started to become something really transformational and really powerful. And it's hard to balance keeping up with that to really be understanding where they might be in the next five years or 10 years , um, and run a company. So running a company has lots of details around a bureaucracy that I built actually, that , uh, is, is quite at odds with the, the , it takes a lot of time. And even just if someone asks you for a decision, it's still gotta be based on lots of time spent researching around that and understanding , uh, you know, all the factors for that decision. So it was like one or the other, really. I was very frustrated trying to do both. And so I realized I should go with where my heart is. And that's where Moodle, how Moodle started in the beginning was me having a very similar phase of my life where I , I had all my time to research and experiment and, and innovate.

Well, I know you are someone who doesn't like to talk about titles, but you are of course a recipient of four doctorates, which is very impressive. Three of those , uh, honorary doctorates from universities, I think in Belgium, Spain, and Greece. When you started Moodle, I know that you are inspired by the opportunities the internet provided to deliver quality education no matter who you are or where you are. And indeed, of course, you've remained committed to , um, Moodle as an open source platform. You've already mentioned ai. Do you see AI providing the same opportunities that the internet provided at the beginning of your journey many years ago?

Yes, indeed. I , I , it's a , it's a wave of technology bigger potentially than the internet. Uh, the internet was or is a connecting technology, enabling technology that we've all gotten very used to carrying around these supercomputers that connect us with the world and being able to answer any question within seconds. And these were not really something that everyone had only a couple of decades ago. And as it came in, there was huge opportunities around education. It was obviously we were gonna be able to , uh, change the way education worked in a, in a significant way. And that has happened. A along with that came lots of downsides with the internet. I think we've launched it upon our society a bit experimentally. Uh, we've allowed social media companies to grow and do whatever they want, and their algorithms optimized for advertising has led us to a place where a lot of social media is kind of really bad for us, and there's lots of people addicted to it, even though the social media, it seems to be actually making social stuff happen less , there's a lot more polarization and filtering. So there's a lot of bad sides to it as well. But also the immense upsides that you can study, learn anything anytime you want, or communicate with anybody anytime . And all of that is , uh, incredible. So now with ai, we have exactly the same kind of situation with a , a new technology. We've invented brains, electronic brains. They don't do everything a human brain does, but they do a lot of the main thing we use for work, which is taking lots of information and producing new language from that. And language can be words, it can be images and all these things. And so you find that we've incre we've created this incredible labor replacing technology that again, promises enormous, amazing , uh, possibilities for education in that you can suddenly have one tutor per person, or many tutors per person. You can have , uh, instant access to not only raw information on the internet, but information that has been thought through and collated and customized for you by an agent. You can have AI's working whole jobs, replacing whole, whole jobs or , or augmenting you with a whole team of effective team members that will help you do things. But of course, there is the downside, and the downside is centralizing of control. The, the increased , uh, data and privacy issues , um, both in the training of these systems and in the use of these systems. And then there's, what I've been most worried by most recently is recently this last week, is this feeling that the , the quality of AI tends to push everyone towards the median, towards the mean. Uh, you have this averaging of society that happens in ai and you kind of get this sort of dumbing down of things in a way. There's less of this kind of more chaotic human people living on the edge, inventors, crackpots, doing weird stuff that is part of the richness of tapestry. And what we're doing is kind of mainstreaming everything. And I , you know, wonder what the effect that will have long term . So there's a lot of potential downsides. And so what I find myself doing is like tackling, trying to understand that , that topic, like where does AI really work? How can we maximize its positives while minimizing the negatives? Yeah, it's a huge job. The whole world is wrestling with it. And , uh, along with myself.

Yeah. In the context of maximizing its positives, do you think it has the capacity to solve some of the world's bigger problems? Moodle has always been motivated to su support the un sustainability goals. Do you think AI can contribute to solving difficult problems?

Yeah, so in the past we would talk about the sustainable development goals as things that people should learn. People should develop capabilities in order to solve these problems. These, the big agenda of , of the problem crises that we face. Now we can train machines to tackle some of these problems. And I, I , there is enormous potential with AI as part of scientific endeavor to create solutions and so on. But , uh, actually implementing those, so solutions is 90% people. Uh, people are not gonna change their behaviors or , uh, habits without a lot of, a lot of effort. There are obviously lots of information about how we can improve climate change, does it actually happen? That's up to governments and , uh, people who vote for them and all of that. So they're much bigger problems to solve, much bigger , uh, pushes to solve those big problems than just technology or just AI that helps find solutions. But I'm confident that it will help. And I think there's a lot of , um, potential for science to advance faster and , and have solutions coming out that , that we can at least then try and implement. So yeah, I'm , I'm positive about that. I , I'm worried with education that a lot of people are seeing AI as a replacement for lots of processes when there's no need for that. Because if a human being is really going to learn something, this brain of theirs needs to do the work to learn it. Yes, you can replace a teacher with an AI teacher. Yes, you could replace a lot of the student work with an AI that does that work for them. And in the end, you just haveis working with ais and nobody's learning anything to actually learn. You do need to put your brain through some stress, like you need to make it work. Uh, you need to actually burn calories and energy and your own little data center and making your brain do that work. 'cause it's the only way neurons and synapses are gonna restructure and you're actually gonna learn anything. So that's, that's really critical. And it's , um, keeping people at the center of education is definitely a big part of what I'm trying to do as we push forward the technology

Absolutely. Augmenting the capacity for humans to learn, you know, creating efficiencies in, in many ways, I think in some ways is what you're saying. Yeah , absolutely . So how does this relate back to your role overseeing research and development or the research lab, I believe you are calling it at Moodle? Yeah ,

So the Moodle Research Lab is , uh, completely new now. We have a history of , uh, research around Moodle , uh, various things that have happened in the past, but we've never had a team that was dedicated to research before. And so that's what I'm building up now. Um, it's new for me as well. So I'm also learning what it takes for this role. I'm getting a lot of advice and looking at what other research organizations do, particularly when the research organization is embedded within a, a product organization that's a little different from, let's say, academic research at a university or, you know, pure research in other forms. This is a bit more product focused , a bit more practical. We're , we're about making things that, that , uh, improve the world. So designing that research process is something I'm working on, working out what my job is, working out what my team is , uh, working out how we can be efficient as well, because we are not that, we're not huge. We don't have , uh, millions of , millions of dollars to throw at this and hundreds of people. We're a small team. We are obviously going to be using AI as much as possible. So a lot of it is about tooling up for research, building the tools we need to do the research. It's all very, it's all very good and interesting and , um, trying to do it in a way that we have lessons. We have a will have a regular feed of, of output that will help everyone in the Moodle community as well. When we build tools to help us with the research or to help us at Moodle, they'll be open source tools that other organizations might use for their research as well. And trying to work in that way, a very open, transparent way.

That's great. So it's as much as utilizing the tools and in turn , uh, helping other people utilize tools in working towards a long view for Moodle, I guess you'd say. Um, what do you think Moodle as a learning management system and a platform, what do we need to adapt to?

The whole concept of a learning management system is definitely something that's up for discussion. I, I think , uh, you know , Moodle itself does need to evolve. As I said, though, there are certain processes the brain needs to do, and there are things we've been working on for the last 20 years. A lot of that still makes a lot of sense. You will, the whole idea of collaborative learning that Moodle is built upon is as relevant today as ever. And it's really important that we build educational experiences for people using these tools with a , a real thinking about what it is that the students are being required to do so that they're learning. And a lot of that actually means keeping AI out of the process. You don't want AI jumping in and giving you answers. You know, if someone's asked you to work on something and you get a a bot to do it for you, you've learned nothing. Yes, you've, you've got a hundred percent for your assignment, but you've learned nothing. It's actually often about keeping AI out of things. And what we are looking at is where AI will help the most is very likely to be on the admin tasks, the drudgery, the , the , I need to click on something a hundred times to process a , a list, or I need to , um, uh, you know, reset my course at the end of the year, or , um, how can I improve this course based on the feedback I got this year from my students? That is a kind of a, you know, you'll be, might be sitting there for weeks, reformatting your course, maybe an AI could help you do that more easily. You could just tell it what to do and it would do, do the actual work for you, and that's okay. There's still a human in the loop. So it's, it's , uh, so I think Moodle does need to use and leverage AI a lot. I'm , I don't think that we need to create AI tutors that are doing all the teaching, and I, I don't think we need to give students so many AI tools that they're not learning either. So keeping it around the admins seems like a fairly safe first bet . Um, that, and there is a survey out right now actually on moodle.org in our research lab , uh, where we're asking people to look through lists of admin tasks that we've collected for students, for educators and administrators , um, and prioritize for us what they think are the most, will be the most impactful , uh, admin tasks to , uh, add AI to .

Yeah, that really touches on that concept of efficiency. That makes a lot of sense to me. Um, from the learner's perspective, do you see , um, you know, any particular tools or developments that will assist their, their learning basically?

Yeah , so I think a lot of us are , are playing with chat GPTs or, or some other , um, ais right now. And it's important to look at your own experience and think what is working, what isn't working. We, we do need to do more research and collect those opinions. We have been doing some surveys already and we'll continue to do more. But , uh, one really big use case for me is , uh, that when you need to look up something quite recent or you want a summary of things, what , what's going on in a field or what's going on in a particular topic, there are some ais like perplexity ai, it's a commercial service, but I think it's, I think it's got a free , uh, layer as well , um, that focus on looking up references. So they'll actually go out to the web and search it and find lots of stuff and bring it back and collate it into a summary of that topic. So that'll, that is a very useful thing. If you're trying to learn something, you really want the latest information about it or the latest expressions about it, that, that's quite useful, that saves you a lot of time , um, that you would've spent maybe in the only a few short years ago doing Google searches and pa you manually through lots and lots of documents to extract those nuggets yourself. So yeah, that's one example. A second good example, if I may, is , uh, transforming content from one form to another. If you have a very long document, you can produce a short summary or if you have to get the bullet points out of that, if you have a piece of text, you can convert it into a audio file like a podcast. And that might be something you can then listen to while you go for your morning walk instead of sitting in front of a , uh, a screen AI transforming from one to another. That really makes it a lot easier to construct your own learning environment if you like. You know, you have to learn all this stuff and so you can bend it and twist it and make it something that's useful for you. I feel like that's a really good thing for students as well.

Yeah, I think that's a form of differentiation, actually accommodating different learning styles and preferences, and I imagine so the learners can, you know, control that themselves. But in addition to that, maybe teachers ultimately or educators can also, you know, utilize AI to differentiate content or deliver content for concepts that a , uh, learner may not have grasped fully in a more efficient and automated way that I would imagine is also potentially a use. Mm-Hmm .

Yeah, absolutely. Um, it's a nice use case of taking a large piece of con content you needed to consume. Uh, it could be a video you had to watch or something you had to read, and you can generate quick quiz questions from that, like a little , um, uh, flashcards to test yourself, you know, test your comprehension of that. And that's something an AI can do quite easily, quite well.

Okay. So what about your current, what are you currently working on in the research lab?

So one of the , one of the things we do is help gather information to inform the LMS development, but that is not my main focus. The , the product teams have their own UX research, their own researchers there that are working on kind of incremental improvements of Moodle. There is also a project in LMS to add an AI subsystem so that plugins can be easily added to Moodle to add AI features. And I think the whole community will be adding little things here and there as we are all experimenting. This is the beauty of the open source ecosystem. So I'm enabling that and helping that and supporting that where I can. But there's two big development projects that I think are really important and they overlap somewhat that are prototypes that I really want to push to demonstrate something beyond what we have now. The first one is a coach. So I think , um, AI has the potential to be a really great coach, and people who get human coaches , uh, usually report that it's really good for them , it's good for their careers or, you know, to support them in life. Coaching is something we know works, but most people can't afford coaches. But I think a coach that was on your own, that , that you had for yourself that was with you 24 hours a day for your whole life would be a really good application of ai. My hypothesis here is that to make it work really well, it needs to know you really well. It needs to know your goals, your dreams, your, your failures, your , uh, what you said to your friends yesterday, what, you know, the , say I'm doing a podcast with somebody, it needs to know what I said. It needs to know a lot about you, I think. And I think if , if an AI had all of that information, like all the cloud of your life, all of your digital exhaust, basically it would, could be a very effective coach because it would know where you want to go, knows who you are, and it can cross that gap and, and help nudge you towards whatever it is you're trying to do. The problem is that a lot of people are thinking about these types of assistance, but usually they're the big tech companies. And the big tech companies are saying yes, you know, Google and Microsoft , uh, and OpenAI will be saying, yes, give us all of your information so that we can give you this lovely service. And that scares me that so much of this private information would be in the hands of these big tech companies who have demonstrated before that they are very happy to , uh, use the data. They have data is the new oil and so on. Even now they have, and now they'll have even massive ais to look at and process that data. And I think there's a lot of, a lot of concern about the potential influence they might have on society by providing assistance that assist you. And they're centralized in this way. So what I want to work on, and what I am working on is a coach that has all that information, but it lives on your phone , 100% contained on your phone. It's not up in the cloud, it's running on your own device. It has access to all your messages and mail and, and microphones as well if you want to give it access. So you can say, oh, you know, listen in on this conversation and just store that as text in your database as well. Um, the memory would be sitting in your control and you could change the brain. So add new learning , uh, new language models, sort of upgrade the thinking power of your device as your device improves and as the language models improve, you can update that. But then there's a memory as well, like a database of all this digital stuff that you've created that is accumulating over time, and that's living on your device as well. And maybe you get a new device and you just copy across your brain , uh, your memory, you upgrade the brain and you know, your coach continues on and it will keep growing with you throughout your , throughout your life. So the main reason, that's a general purpose thing, but where it applies to education is I think a lot of what life will be is gonna involve education. So this thing will be helping you find those educational opportunities and connecting you up to them. And so that's like a big-ish project . It's a big-ish project, but I've had some success with some prototyping already. I will be pursuing that as one of our, our bigger , bigger projects in the research

Area . That sounds very exciting. And that database is a , as a matter of interest, is that, so you say that's stored on your device or on a server of your choosing? Is that the inference

On your device Yeah, on your device? Yeah . Devices now easily handle gigabytes of data and , um, we're talking mostly text. Everything will will be crunched down into text and stored very efficiently. So you would have, if you think of a language model itself, some of them are only a few gigabytes and they basically appear to contain all of human knowledge. Like they have incredible amount of information encapsulated in a few gigabytes. So I feel like a a a memory can be quite effectively done too .

That makes sense. Everything

Is, everything's improving and it needs to be in your control. This is very critical to me. So this open, open ed tech needs to give you that trust. You need to trust it. And it's very hard to trust big tech, which is driven by, I mean, just look at the behavior of Microsoft, for example, the way they jumped into ai and they've put copilot buttons on their keyboards and they're building it into everything, and they're looking to just scoop up all the data of the world and, and have it come through their big central powerful system. Now they're spending a hundred billion dollars building data centers across the US that will run the future. You know, they're already thinking about GPT seven and things like that. That's like a almost maniacal , uh, focus on control that, you know, if it was a science fiction movie and you heard about some evil villain who was building a hundred billion dollar data center in the desert to <laugh> to uh, you know, do everything in the world, it , it's kind of scary. And I I, there's no need for that. There's a , there is also a big push amongst people for a decentralized approach, having everyone having AI assisting them, and that's a bit more human right. We all have a lot more autonomy.

Yeah. It's a good segue to open Ed Tech Martin, because I know that's important to you and it is a part of your focus with the research lab. Can you explain that connection? I mean, maybe you have in that, in that context, but a little more understanding. So when ,

When thinking about education, it takes a lot of players to work together around it. We are developing this sort of coach prototype and um , hopefully it works. Uh , it works out. I mean , um, that is something that will be out there. It, it needs to connect to a lot of data. If it's going to organize a course for you and sign you up to something, it probably needs to interact with a Moodle site or some enrollment at some university to, to , to do that. And it, it needs to be able to negotiate. If you are already in , in a course, say you're in a Moodle course , uh, and you are studying away and you want your coach to know what you're doing online there. So Moodle's gonna have to feed information back to your coach. The, the coach might need to perhaps handle some of the admin. It might need to sign you up for things in that Moodle side , or it might need to go and explore what else is going on in there. There's quite a lot of little scenarios where you need to pass information backwards and forwards. And so that's a place for a standard, that's a place for a , a protocol that we could agree on as a education community that all of our tools support these standards and therefore can talk to each other and interoperate that Tide will raise all the boats, will all, all the tools will work better because they'll connect. So that's a key thing about open ed tech , is about how do we build this open framework? How do we recommend certain standards over others? There's one other aspect to open ed tech , which is when a product or an organization making tools conforms to these ideals, so it's open, it uses open standards, and additionally , uh, it listens to educators for its roadmap and it supports open, develop , open source development through the community. If it complies to these things, we can certify it as open ed tech . So to make a bit of a shorthand so that if you understand these principles I'm talking about, you can look around and say, oh, you know, which products are open ed tech and which are not. And there'll be a , a certification stamp on, you know, like an iso uh , certification or something like that. But this is very specific to the open ed tech ideals. And that way you'll be more, it'll be easier to select product to use in education. 'cause you might say, look, I'm only using open ed tech products, which is how I imagine it might work in the future. If they've got an open EdTech stamp, then I know it has a, B, CDE, it has all these features, and so it , it makes life more efficient. So that's the reason for open EdTech. Yeah. And so part of my job at the research lab is , uh, supporting the, the spinning up and getting the, going , getting open, EdTech going , um, and it's , uh, it's , it's improving week by week.

That's exciting. And, you know, that involves working with people in all parts of the world and bringing people into the vision. Is there opportunities for research with external people as part of the research lab? Do you have any scholarship, you know, ideas?

The Moodle research lab area on moodle.org is kind of the hub of where we're doing things. And we have some forums there and some chat groups around ai. And a lot of the Moodle community obviously are very excited about AI in particular and other new technologies too. I mean, let's not forget about augmented reality and so on, actually. So some of the very first , uh, augmented reality glasses are on their way to me as we speak, which I'm, I'm keen to get on. And these are glasses that are very thin. They use waveguide technology , so they're not, it's not a headset, it's actual glasses with a , with screens. It's all early days for that too. But it's also interesting new technology, but particularly ai, it's really the, the , the focus in , uh, in the Moodle research lab area. We have universities and , uh, other people from around the world helping. I hope eventually we'll be able to sponsor scholarships for PhD students or post-grad students to work on some of these issues , uh, with a , a Moodle scholarship. So that's something , uh, we're, we're, we're putting together as well. Otherwise, you know, there is so much activity around the world on these topics. It's a full-time job. Just even to just look at the surface. There's 4,000 papers on AI coming out a month at the moment, like academic papers. Like it's, it's impossible to read it all without ai. You need AI to summarize things and just keeping up with that torrent of development is , uh, insane. So the whole world is in a sense, already working together on a lot of these things. Just ,

Yeah, it's in the spirit of open source, you know, a collaborative approach. Yeah . To, yeah. We'll solve problems more quickly. I recall a quote that you used in a recent interview. You said A golden age for broad learning, a new re renaissance. I love that quote. Yeah . And I thought it would be nice if you could explain , um, what you meant by that.

The best possible scenario is what humanity does with , uh, electronic brains that we've invented. The , the best possible scenario is that we get the robots doing all the work. Most of the work, the, the boring admin, right? The drudgery of, of running a society. So they're out building roads, they're out building buildings, they're running call centers. They're, I don't know , doing management. I mean, I , I actually think some of the high level management roles are easier to replace with ai if society is run by lots of robots. And I say robots because it's not just AI software, it's the hardware robots that are coming as well where AI is now expressed in the real world with real world input and, and humanoid and , and other abilities to manipulate. They'll be cleaning our houses and doing all that stuff if we can have the bots doing all of that stuff. So we are freed up as human beings. We don't have to work for a living anymore. It could be that golden age where we are just freed up to think and discuss and do community stuff. It's hard to think about that because in certainly in all our lifetimes, and for our, at least the last few generations, the economy is completely predicated on the notion that you need to work to survive, right? You need to have a job to get money to pay for things and to help run the economy. And the government prints money to circulate. And the whole economy is like on this entire system. When you have robots doing all the jobs who don't need to be paid, that breaks down obviously entirely. You need to rebuild the whole economy around some sort of a universal basic income. UBI is the phrase there where everybody would get money given to them and it would be a very good salary, and you would not have to work for it. It wouldn't be means tested even like you would just say, let's, for sake of argument, let's say you just get a hundred thousand dollars a year given to you, right? And you can spend that how you're like, what are you gonna do if you're not working for a living anymore and you suddenly don't need to, and you've got that money? Well, I think you'll be learning. I think there'll be a lot more learning. You'll , you can follow your interests and get into niche areas and become an artist, or become an artist and a race car driver, or become a, you know, do all these things you want to do and explore life, right? Without the necessity of having to turn up for work every day . It's not like, we don't know this isn't possible. 'cause if you look at rich people, that's what they, that's what they're doing, right? They don't have to work for a living anymore. And they're having a seem to be having a grand old time out there <laugh> , um, with tons of assistance doing everything for them. I think if we were all in that situation, it could be a very interesting world. It's certainly the only kind of path I'm , I can see that I like. 'cause the other scenarios are quite bad, where you have AI used to enact power over people, and you have an increasing inequality and you have, you know, well look at every dystopic sci-fi movie that's ever been released, <laugh> . Like, there's a lot of bad scenarios that could happen. I think we should be pushing for that good one. And that's why Yeah,

I think think it's a beautiful vision and it's something to, yeah, to , for us all to , uh, aim for. And because it's this concept that actually the ai , uh, and new technology actually helps us become more human and realize our own individual potential, which is, which is of course, what learning and education is all about. So it's a really nice circle. Thank you for that. We come to the end. It goes so quickly having a chat to you , I'm afraid. But I , two things I want to finish with. One is I'm sure people in the community look forward to an opportunity to hear you talk about these things in person. Uh, we just released our dates for Moodle Moot Global this year, and I imagine you will be going to Mexico October 22 to 24. Yeah,

I'm really looking forward to that. I mean, that's our big global conference every year that , uh, we get so many friends and community together from around the world and , uh, always new people as well. So , uh, it's really good place to, and I, I enjoy that so much for those few days to actually seeing everyone face to face and , uh, sharing all these ideas and everyone gets very energized. So , um, looking forward to that in Mexico.

Absolutely en enlightening , that's for sure. Okay, to conclude, you mentioned before that people can become involved and see what you're doing and become part of , uh, surveys and research. Where do they do that, Martin ?

So you wanna go to Moodle, me slash research, Moodle, me, me slash research. Brilliant. And that will take you to the , the course on moodle.org where things are happening and it links to everything else.

Always a pleasure to talk, Martin. Take care.

You too, Abby. Thank you. Alright , see you everyone.

Bye. Bye everyone. Thank you.