I think this is working So, I'll take you take you through some of that history
over the years, I'm, I will, I want to do this to give you kind of visibility of what how we think what we're trying to achieve. help you understand the way we think the way we approach things that we were approached, developing memorize. And I, my hope is that will give us more context, on memorize and hopefully build towards more constructive dialogue that we can have with all of you and that we can. We can all be talking from the same context and understanding where we're going. And so that's the aim is like open up the book, take you through our thinking, so that we have a discussion from that basis. And from that basis, I then want to open up and hear all of your questions and answer them as well as I can. And as I think we've I've said on the forums, I don't have answers to all of the questions and some things we just don't know the answers to yet. But as much as possible, I want to give you the context to how we will be going about answering the questions if I can't give you an answer right now. So I want to start with just a little bit of the distant history because it is important to the context of what we're doing and memorize. And that is why I founded memorize Why did we start the company? So from my perspective, I struggled with language learning at school. I found I could learn to pass exams, but then when I got to the country, I just couldn't speak the language. And I just couldn't understand why this happened. And having to be French having to meet with Spanish and I just, I didn't get it. And then I went and I was I did a master's in psychology at Oxford. And while I was studying that, mainly focused around memory I came up with a theory as to why that was and has to do with the specializations in the hemispheres of the brain. And so I just taught you quickly through this because it's an important bit of background to understand how I think about language learning. So, language tends to be spoken of by psychologists as being a left brain specialization, their focus area, Broca's area in the left brain, that are crucial to language processing. And broadly, they address or they process the meanings of individual words, explicit meanings, and the order in which that said so roughly a kind of grammar that and that's what pretend to teach me teach language lessons, we teach very much to those bits. But do most studies I came across this idea of like, what the right brain role was in language learning. And the right brain has really important roles. It has roles in helping understanding from context, metaphor, understanding of motion, humor, expressing your wishes, pasady, tone, all of these different things and fundamentally relating to other humans through language. And when I looked at those two lists together, and it was like, relating to other people for language, expressing my wishes, understanding humor, on the one side, an explicit word meaning and word order on the other, occurred to me that that was what I wanted was the list of the on the right. That was really what I meant when I said I want to learn a language. I wanted the stuff on the right, the stuff on the left, it's like well, okay, if that's gonna get me there, I'll, I'll take it, but I don't want that on its own. And what I'd been taught at school was just that lift list on the left are just been taught, what are the words What did you put them in? And none of the stuff on the right, and that was kind of left to chance who's left and when I went and spent time going on French exchanges or whatever, in order to fill in that bet on the right. So what I wanted to the theory came up with was actually we just need to very explicitly overweight the teaching to the right side of the brain and teaching those kinds of skills and putting people in the right context to learn those. So I went after graduating from Oxford, I went off to China basically to test out that theory myself and just to see if I could succeed in learning speak a new language. If I consciously tried to teach in this way, and I found that I could, I could do it quite much faster than I thought I was going to be able to do much faster than people told me I was going to be able to do it. And that experience is where the impulse to create memories really came from it was trying to bring that spatially Hemi spatially integrated approach to learning languages to a wider audience. And so that people more people could benefit from from that experience. And the way we talk about that is that there's there's a learn side is broadly the left hemisphere specializations and there's the a mass and communicate which is just processing a lot of language input, and having a go to process to work out how to say things how to try to express yourself on the right brain and says, Learn and immersing communicate and need both.
So that was that was a long ago background and where I where I come from on this. The early journey with memorize, we started with Advil, we want to do both. Let's start with the Learn piece because that's the that's the easy piece. Let's get that in place. And then we can move we can build from there. And so we started out building something that looked a lot like flashcards. It was I mean, I remember the conversation in 2010. Where, with Ed, my co founder, where someone said, case they were first building flashcards and I was like, yep, but what we're building is Novan 32 flashcards. That was the intention from that day, more than 5% It's an important tool, but it's one part of a much of a much bigger thing that we need to build. With. We began by focusing on languages we then the tool was taken up and used by people to learn other things. And we only encourage that activity I helped. I spent hours helping people make art courses
and learning it wasn't as
you born, steering away from the vision that we had overall there was this core community even in at that point, I 2014 that really did enjoy the product, but it was growing like 10 15% a year, which is just it's very slow by Internet standards. And it's slower than it needs to be to be viable. So we we took a long thing then we took a break and we thought what, what are we doing wrong? What do we actually believe in? What are we here to do? And we went back to this Learn immersing communicate. This is what is what we believe is why we set the company up. So then we looked at what is blocking us from doing that right now like why why are we not bringing that to the world. And what we found at that point was that in order to bring these immersive and communicate communicative experiences to into the product, we the the structure we had of allowing anybody to make courses in the database, and the same tests being used against any content that people put in was making it hard for us that was blocking our innovation. And so we made a decision then, what we needed to create was new test types video comprehension, audio ones, to give the full range of immersive experiences. And we couldn't do that without that open structure. So we made the call to move from what memories have been which was using a courses first to memorize courses first. And that was a big change. We we needed to do that we need to create a central database of words and phrases where we create controlled the content and the structure of that content so that we could innovate on the design of tests and learning experiences, because it's in the design of new tests and learning experiences. That we should make something that was radically more effective than what's around the what is available otherwise. So we did that we made that move in between 2014 and 2015. It was a really hard rollout. The existing community broadly really didn't like that we were doing it, downplaying the use of user generated courses up playing the memorize ones. And there were those who felt that we were really betraying What memories stood for, and the memories that they'd been part of building and it was really, really hard. But what we what we found was that move unlocked really really rapid growth. So as the cost of two years after that the numbers of new people move from just to put some numbers on it move from no hundreds of 1000s of new people coming into us. So using memorize every month to over a million with no marketing budget, having been very, very flat before that was really spectacular change. And that was driven by really focusing on what we we believe that is differentiated about the way we want to have the kind of language learning experience that we want to create. And it's connection to real people. It's, it's immersive, it's all of these things, all of those things that caused us to set the company up in the first place. That also meant because we were getting such growth that we could carry on supporting the old legacy UGC user made courses product. It gave us that optionality. If we hadn't made this move, we wouldn't have had that optionality we'd have run out of money, and the company would have died and the website would have gone down. We made that move. It allowed us to continue to support it.
That period of growth continued, really up to the end of 2018. And that period of rapidly rolling out these new features and making these memorize causes better. And then we began to run into some troubles and I'm being I'm being like, really open here. I wouldn't, wouldn't ideally say this to a press release. But I am just going to share as much as I can with you to give you because this is this is what affects how we think and the decisions that we make. So I want to be as as open as we can. So at that point, we had lots of ambitious plans for the kinds of experiences that we wanted to roll out. And we were soon to be able to roll out now that we controlled the courses and that in taking control of the course structure. We gave ourselves this period where we ran really fast, made huge impact through really quickly. But then we started to run into the sand again. And what we found was that we were getting swamped that even though we now control the courses and the course structure. Still there was a lot of technical changes that we needed to make in order to get the new experiences live. They just it just wasn't a clean enough break. Wasn't a clean enough ring fenced a piece of dentistry. And it meant that we kept having to deal with complex edge cases in anything that we were rolling out that technical and product debt as we call it has just stifled the speed of movement and that's meant that effective new features we've released a couple in the last year. The comprehension tests, the Whiz quiz and the communicate feature on mobiles on iOS are examples that have been really well received and had a really positive impact, but it's nowhere near the level at which we need to do it. And this is a really big problem. Because tech products need to move fast. The world moves quickly people catch up, they copy what's happening. And and they the market changes because of that. And if we don't keep improving the product and staying out in front and bringing new things. The memorize user base will not continue to grow. It'll shrink and if we start shrinking, we will not have the funds to support even the current product. So we do need to find ways to keep growing if we get to a certain critical mass. We will be okay. But we're not at that level yet. We need to grow from where we are in order to have that long term certainty to saying staying still keeping the Legacy product as it is and prioritizing that is not an economically viable option. We have got a very well evidenced roadmap of of development of features that we have a high degree of confidence can work can be highly effective in helping people learn languages. We have a track record in the past of creating features that we believe in and that deliver a great experience. We have a recent track record of not shipping those features fast enough and that is what we we as a company urgently urgently need to change. So in the middle of this year, we made a decision that we needed to change approach. We been trying the same thing and the same way of approaching it and kept running into the same Soundtrap. So the way we see it is in the last few years we have been slowed down because as we've rolled out the new features we have tried to solve for every possible problem in education before we move forward. And in a code base as complex as ours is. With that legacy of 12 years building on layer on layer that has effectively meant we've round to a halt. To change that we need to change the approach. So rather than minimizing risk, we need to maximize for speed of movement towards that best possible possible outcome. We need to get where we're going as fast as we can. And what that means is that we do need to make some decisions. That will be a step backwards before they're to step forwards and we do that knowing that that's a risk. We made this decision. Knowing that this is a risk. It means that we're going to make changes that will break things. That is a fact it's not our aim. Our aim is to get to the goal and the experience we're trying to create as fast as possible. But it is a fact and it's a side effect. And I want to we know that you all know that and I just want to be clear that that is a conscious judgment that we need to make.
To but what will happen if we don't do this? And this is very much the discussion we've had internally like what will happen if we don't decide to change our way of operating or make this shift when if we don't happen. If we don't do this, if we don't continue to innovate and build out that vision, build out the features that we have highly evidenced that will work and that will help people learn languages faster, remain two things. First of all, it'll mean that we will not make language learning genuinely easier for most people in the world, but fail to do the reason that we set this company up in the first place. We won't make that happen. But more pertinently, if we fail to make that progress, memorise will shrink. And if it shrinks and the user base shrinks and the revenues shrink, then we lose financial control of the company. And ultimately, although we're a long way from this, if we fail to make progress, then the memory site will be closed down. Like that's just the way it the way it will go. And staying still is not a viable option. We need to make that progress and we're not prepared to risk losing that financial control. We're in a good position at the moment. We need to move faster in order to stay in a good position. We can't take that risk. We can't hit the risk for us. We can't take it for any of you either. Because that's not what any of you want. Now this is a kind of a nuanced argument. And I'm just pouring out a lot of ideas here. I'm sure there will be a lot of questions about what I've said. And I also want to be really sure that we don't that no, no one takes any sentences on their own or individual things that I said and extrapolate something from that that I didn't intend to mean. I'm sure there are lots of ways I'm saying this that could be misunderstood. So let's please ask questions. If there's anything I've said that you think might be taken in different ways or you find shocking or like I might be sending something wrong, please let me know. So we can talk that through and see if we can see if I can correct that. But I just want to kind of run over a summary of it again and then I'll open it up to questions. So we have a vision of what we want to build that is integrating his left and right hemisphere learning. We are aiming to build that which is what we've been aiming to build from the beginning. And that is what we believe gives people what they need. In order to really learn to speak a language. Notice not a part of it. That experience will include as you're having said this, so let's just be really explicit on this. They'll include tools that let learners make their own lists of words to learn. And the list that they make will have content for those words from media from words being used in context, in videos in audio. It'll also contain tools for you to auto generate your word list, from films from media, from YouTube films from podcasts, from conversations to save you time, that is absolutely a part of what we are building. We know a lot of evidence and I'm sure many of you have the same experience that most language learners keep personal list of vocab that they want to learn. I certainly do it. And I know most of the team knows. We're making tools to make that much easier for you and more effective. That's a really important thing to us. Building that reality is something that we are heavily diehard committed to delivering. That is what is our on our roadmap, what we are thinking about every night, every day, when we wake up in the middle of the night. We do have as we're moving towards getting to building that we have a lot of tech and product debt that has built up over the last 12 years. And that has slowed us down in particular over the last two to three years. We can't continue to stagnate development. We need to innovate in order to reach the goal that says both current users and future users as well
as in order to do that, we've got to take some tough decisions. We have tried really hard to do this over the last years without taking tough decisions. We couldn't make it work. And I'm sure and I know because I've read it on the forums, that many people on the forums question my competence in doing this. And I don't have an answer to that. Except to say that I'm not the one only one that's been doing this. We've had many highly skilled people working on this problem. How can we move fast in that direction without getting stuck by the technical debt? We couldn't manage it. Maybe other people could. But were the people here and we have tried very, very hard right now. The judgment I'm making is at the risk of continuing that approach is just too high. If we don't make progress on building out the vision for another two years, then the company will be under really serious threat. And the dream will be over for everyone. And I'm not prepared to let that happen. No no Memrise is prepared to let that happen. So that is why we need to do something we need to make some tough decisions. Now. Know that as we're making those tough decisions. Know that the place we're aiming to get to is a place where users can make a list of words, users can make, use memes and create notes on individual words. as two separate things really in my mind, users can experience a range of media together with personal native speaker advice and they can generate word lists, from the conversations from watching films from all of this and that it where it is possible. To keep the legacy systems running in parallel. We will we're not going to break things deliberately. It is not an aim. It's only if we can't work out how to get round. So the user made courses are as an example of that. The way that we're managing with Udemy courses is we're first building the ability to make for users to make lists of words and courses. Within memorize within the memorize dictionary having the benefit of all of the media content and different test types and comprehension tests and so on that we have there. First we're doing that we're leaving the user a course is running in parallel. We then we will move over as much as is possible and wanted to content from the user may cause legacies and courses into the new system. When we've done that, we will look at whether we still it makes sense to continue supporting the legacy. Causes is my causes. Non language causes as I said on the forums are not part of that future vision. So non language causes cannot be moved over to the new thing. That doesn't mean when definitely gonna kill that product. We're not definitely going to retire it. That's not a decision that we will make unless we absolutely have to. If we can get the focus on the main memorize product and get that moving again and shipping features quickly. I'm making making progress in the direction we need to then we have the resources to be able to continue to support that. And we absolutely
we would probably make the decision in that direction to continue to support it, because it doesn't make sense for us not we would only not support it if we had no choice. And so that's that's how we think about this. I can't give a firm answer. We're going to support it or we're not going to support it because we don't know. But our intention is we are building this vision of the main memories product as a language learning product. Most of the language courses can be moved over into that the ones that are non language courses. We will see if we have the resources to continue supporting it. We will continue supporting it. If we can't then we won't, but that'll be because we can't. However, in the case of memes, we can't work out a way to hold them in place while we make this shift. I know that this is painful, and I know it must be hard to believe from the outside that the tech can be that difficult. I personally have made 1000s of memes learn Chinese, I don't want to lose them. The intention is as I've said that memes will be coming back into the product, as I said on the forums and be misquoted a few times. I can't promise that or a timeline for that. But I can promise that our principles of how we believe languages should be taught I've shared those 12 principles of learning must communicate. Those are what we are operating to. Memes are the best way I know of to serve. Learn principle number two encoding written memories richly that I can promise I cannot promise a timeline for when they'll come back. But it is our intention. It's absolutely our intention. And I'd say I think the summary point there is where it is possible. We are keeping legacy systems in place and just moving to the new ones where it isn't possible isn't possible. And we but whereas over the last two years where hasn't been possible. We have sacrificed moving fast and making new features. Now we're shifting and where it isn't possible to do both. We are prioritizing moving fast in building the new features that we need to build in order to bring this vision to reality over breaking things in the short term. And we will fix those things as soon as we can. So that was a lot of talking. I think that probably covers everything that I've got to say so how early or last year. Have you put in how we're going to ask questions or what the best way to manage this is is that in the
yes. Is slack yeah, there's
some slack.
Inside do right we have slide do Where's a whole bunch of questions. So the polling questions. The first one is concerned about the future of UGC is my right processes, all the samples that are totally useless.
So first of all, I'm just going to introduce Steve, Steve turn your camera on for everyone. Except your boss or something. Steve is Steve's I'll see but the money so so Yeah, Steve will presumably you can you can on some of these questions they'll take most of them say people have expressed concern about future view disease. So let's put this into language and non language. Causes. I'm trying to see where you are on my screen as your camera on Jonas. Jonas.
Yes. Hi, Ivan. Amazing. Hi.
Nice to most to put a face to the name and yeah, I've enjoyed our exchanges on the forums and nice to see you so you are, let's put into two buckets. We've got language courses and non language courses. So within language courses, broadly what you're saying is right, if people want those courses to be moved over if they're still active and being used. We will make every effort we possibly can to move them over. There may be some things I can't think of them there maybe there'll be there will be use cases it will be difficult. So for example, if people are set up their databases to test in peculiar directions because they want to be tested on case endings for protect for words or something like that. And it's a particular use of it, that the system, the new system doesn't support that there may be ones that will be hard for us to move over. So there may be some courses that are useful, but we can't work out how to move over any courses that are there for learning words and phrases. We will should be moved should be moved over another subdivision there. So if I spoken to languages in which language pairs in which we have official memorized courses, and therefore we will initially be building these centralized databases. So that's where we will start. We then it is a very strong intention to go back like in the early days of memories, one of the most exciting things was the variety of languages that people were learning or memorize. One of the disadvantages that we had, in moving to memorize made courses was, although it drove great growth, it did necessarily reduce the number of course, languages that we could support. However, we believe in the new system, that's actually going to be much easier for us to solve and cede these dictionaries. We would, the intention is that for languages that we don't currently teach or memorize, so there are language courses on user made UGC, memorize, but we don't currently support them. We would be looking at how we can create a space for those in the new memories and look at whether our good courses are can we immediately see the dictionary with content? So for language courses, short answer, before we get too involved is yes, we will. We shouldn't be moving over everything that is good, except if it's courses that are designed specifically for for drilling on grammar structures in a way that doesn't fit into the new system. I
think some people are concerned about definition of being used because especially after the decision to retire MEMS, another consideration that they are not used. But in fact they're actually used. So I guess some people are concerned about Yeah, will you maybe
I guess a way of thinking what, what I mean is useful if one person is using it, and they're using it by going back to it once a year because it's really important to them, you know, we'll let them know that when they need to tell us that they want it moved over and then we can move it over. I believe that is the approach we'll take to it. And I it there will be a limit. You know, there are millions and millions of courses on memorize. We don't want to be doing it manually on each one. There are more probably more courses and there are active users each month. And we will need to make some pragmatic decisions. But if you have a course that you want to be moved over, there's no reason to believe that I think we might need to mute you.
Excuse me can I may i? So, because people may have some some worries because some minority languages that may have may not have sub active learners so but they are extremely, extremely valuable to memorize especially consider that manifesto alone is upon majority languages. So I really hope that memorize can keep the pot and night oriented languages even though they may not appear to be via very active, extremely valuable to the community.
I agree with you. They are important to the memorize team as well. And me personally. I totally agree with you. And yes, that's the point that I was probably at confusedly trying to make just now is that I believe that the new system will actually make it easier for us to support those languages. Well, we can we can take the content over to the new system. And then we can look out ways of supporting the growth of that those dictionaries the themselves knowing that they're powering a whole range of different experiences. When when they're just individual courses. It's hard to get a list that are not linked to each other. It's hard to get momentum behind improving them when it's a centralized database, which actually, this is the data structure that we used to have memorize right back in 2010 2011. And we moved in 2012 to a data structure where courses were not linked. But prior to that we had a structure where there was a wiki for each language, which got better and we'd have multiple people working volunteers working on improving each of those Wikis for each language. And that's much more similar to this data structure that we're moving back to. And it means that it actually becomes much much easier to get a small group of people creating a course in Masami and making it really good really quickly. So I can't promise that we're going to put a huge resources into that immediately. It is definitely in our minds and something that we discuss as a benefit of this data structure is that we can support those languages better. And back to my point about retiring the UGC we won't the legacy courses. When we're not going to just kill it off immediately. We will run it in parallel and do everything we can to move everything valuable over and I totally agree with you on the value of those minority and niche languages
in economics and Sundays, so am i right to say that if a cause has a user has only one user then is considered used and useful to be migrated to a new platform?
That's,
again, when I try and try and take a run at this, I've had the advantage of being able to read all the questions while you've been talking. And I think I can stick the landing answer and a whole bunch of questions all at once. You know, including this. So what we are doing here is unraveling a whole lot of legacy code and complexity. That doesn't help us accomplish all of our goals and your goals. And so what we are doing is we are stripping back down to making sure that we first get our dictionaries correct and in order i We need know the parent of hola and hello or Allah and Bonjour. To be right and to be accurate, right we cannot have a dictionary that says Allah means bathroom. Right? So first and foremost, we have to strip back to the dictionary to make sure that we have an accurate high quality dictionary. And then to that we attach all of the content that makes memorize memorize or learn with locals and local people saying the words and local people using the words in different contexts surrounded by other words that you don't know. And we need to build that up as well in order to realize our vision to do that district back down and to get that built right. We got to start with those basics. And in order to start with those basics, we can't bring everything over including MEMS to start. So we're going to start building from the ground up, you know, those basics. Once we do this and do this well, then we'll start to allow people to make lists of their own words like they can in the UGC world today. And when they do that, not only does their list of words become their own unique list of words, but it is actually attended with all that great content. When you put a lot in your list, the 15 videos that we have of different people in different parts of the words world saying Allah comes with it. And those videos have immersive content that uses the word Hola, and many other things comes with it. So imagine being able to make your actual UGC list, but complete with all of the content that we've created attached to it. And so in order to make that happen, we again need to start fresh and clean to make that happen. So we're going to do that. And when we do do it, we can then go over to the UGC lists that you've already created, and start to bring them over and all of a sudden your lists of words that you've had for so long. Now get appended with all of the amazing content that we have in our system, your lists will instantly become more awesome. Right? It may be that we can't bring over the memes portion. At the very, very beginning of these lists. And we're not going to deprecate the server where these memes and and your list live currently, until we are satisfied that we've gotten enough useful content over into the new system and everybody is happy. So
so that is the general flow of what is happening. There's another reason why we're doing this. You all have proven that the system is much much better when you're all out there creating content, right? And so if we simplify things into this new architecture, it will become much much easier for you all to create content. Not only can you create new entries into the dictionary, you can create your own learn with locals. If you wanted to. If you've wanted to turn your iPhone on yourself and speak words in you know, different languages, it can then get appended to the words in your dictionary, and then it can get shared with everybody, right? And so the things that you do can get shared with everybody. So these are the exciting things that we're doing in order to immerse people and help them communicate. And we're building you know, from a clean slate, and then we'll start to move things over. Along the way. Some things won't make it. There may be certain dictionary words that are inaccurate in some user, UGC. Eric, right. We're not going to bring over somebody who made a word pair of hola and bathroom, right? That just can't be so so we are building something exciting. We are going to be porting things over and when we build this, your lists will be more exciting because of the content we can add to it. Your list will now enable you to add your own content to it eventually. And we can also start to use this to build other languages but we have to strip some some things out first, before we start doing this.
Thank you, Steve. Yep. Awesome and held on so I'm going to try and rattle quite quickly down a few because there are a bunch of questions there. And I want to make sure that we we cover as many as we can. So when you fix the app, so that UGC is searchable within the ATS. This I believe have benefit memories financial. I think context for this is we go back to what I was saying earlier when we we create took control of memorize courses and of the all of the content. That is what led to to rapid growth and more value to users in general. Oh, was I muted? That was weird.
That's my fault. Sorry about that.
No worries. I wouldn't take offense. So where was that that unlocked growth, having that richer experience where there were videos where there was audio where there was high quality content? We hid my user made courses, because they negatively impacted the experience of users in general. And I know that for some users who started who were using memorizing a particular way to make lists of words that made the memorize apps less useful. But we we did it because it made sense for learning outcomes in aggregate, and it made more money. We wouldn't have done it if it was a very expensive thing to do that way. But that is what unlocked the right so I think that kind of that's why we're not going to be doing for the current user make courses. But as Steve said, we are building the ability so that user made courses will be in the app and that gets to the same situation, but through a different route. Okay, when a smaller languages fan will
be searchable, right? And once we have it in the app, you will be able to so besides will be able to search by courses made by Jonas,
right? So we're not doing it just by putting the current ones back in the app. We're doing it by creating the functionality within the apps
for those I'm not sure if I got your point correctly. But in I say, I'm saying that in some ways, making making ugcs searchable within the apps will somehow negatively influence
Yeah, so. So the current user made courses, making those available in the app, but when they were available in the app, and we took them out that benefited our users in aggregate. And so putting putting those back, there are two things that we could do. One is just putting them back in the current form. But we wouldn't do that. Because the experience there is not part of the vision that we're building. And it had a negative impact when those were all available. So we're not going to do that. But we are going to create the ability for all of the lists of words that use as a making to be in the app for users to make lists of words and have them in the app for user to be able to search those lists of words in the app. It's just that we have the legacy system here and we have the new system. We are not putting the legacy system back into the apps that didn't serve as well in the past. And we have no reason to believe it would now but what we are going to do is create within the apps a way to have the same functionality to be able to create lists of words to be able to learn those with Steve says, as Steve says, With all of that associated media and different test dates and so on, and to be able to search for all of those within the apps. So that functionality is going to be there. But it's not just taking what you currently can search for on the website and making yourself searchable on the app.
I still don't quite understand why is the negative impacts to do it in the current
I mean is do
you want me to take alright, I mean I can tell you from a straight up metric standpoint, I the best metric that we have as to whether somebody is successfully acquiring a language is how much time on task that they spent. When we have our system filled with all sorts of content beyond the memorized curated content at this point in time, that statistic goes down. People spend less time in the app and that is not good for trying to acquire a language. So when when we have objective measures like that, there is not a subjective decision about any one person's course is not particularly good. There's many many good courses out there. But in aggregate on the whole, we have objective measures that say people will spend less time acquiring a language in our system with this content in there versus with this content out of there. So it's a straight up
number. And we don't know why that is. We know what the what the data is. We have theories on it. And I know that people will even now be saying Yeah, but you're Miss reading that data because that could be because of x y and that said, that's true. But we have gone very deep. We wouldn't do this in this from our very best analysis. We believe it to be a negative. And then secondly, what we are trying to build is all of that benefit, but in a much richer way. That is easier for people to make courses easier for for people to make courses that we know are better. And those are the tools that we're building and we want to get there as fast as possible and not take any kinks in the road. Does that does that help?
Yeah, it kind of a question. But I can I can I quickly go back to Steve's point about making dictionaries of language pairs. I appreciate the intention and effort that they're going to put into making dictionaries to make sure that word pairs are accurate. But I I think I but I have a concern about minority languages because for example, I can't I I thought if you have if you have the resources to make some rare language pairs for example, Burmese to Malay, so So there are causes of of those language pairs, but I doubt if you have the resource to make taxonomies out of those rare pairs so
so so that's a fair point to two parts to the answer. The first part is that course of Malay words that is sitting out there today is not going to disappear off of the servers because we're not going to deprecate those servers until after we reach some point in time where everybody is satisfied with the new experience. So it's not going away anytime in the near future. And not without a whole lot of warning. So no, no concern there. That's one part of the answer. The other part of the answer is I have a very interesting and eclectic background in the machine learning world and a previous company we acquired a business that has been creating word pairs using machine learning in order to comb the web and and really do a good job at extracting language pairs phrase pairs and so on. So while you are right about the scalability challenge, machines are going to come to the rescue for a very, very large quantity of what needs to be done out there and take the scalability problem out of the equation. So your point is a good one. We're not going to solve it tomorrow. But we're not going to deprecate the server that holds it tomorrow. So the the current UGC courses safe and the dictionaries that we want to build do have the chance to benefit from deep learning out there
if we don't have enough
Yes, I'm just really I'm really I'm really aware that there's there's a lot of people and I just want to make sure we get through some of the other questions. I guess another question, but let's just make sure smaller languages were spoken to that one. Will you provide a let's come back to I can stay on for longer so I am happy to talk past the hour but I just some people are gonna need to leave a five and I just
and I will be one of them. So
yeah, so it's gonna rattle through the top ones on slider here because there's just a whole list of them and I want to see how far down I can get there. And yeah, you're gonna see if you're around at the end. I'll be happy to come back to that. But will you provide ways for those of us who spend massive amounts of unpaid time creating content to export and save what we created, including audio, memes, etc. This this question seems to take us a predicate that we're turning something off, as Steve says, We're not turning anything off until people are satisfied. There are already as someone in the community has created saying for downloading memes. We're actively looking at whether we can make that better and help support that.
Mine and part of what's behind it is that one of the courses I've spent the most time on is not a language course. So you've just told me we'll get rid of it at some point.
I'm really, really glad that you've asked that question. That's not what I intended to say. What I intended to say was the We Are the language courses don't have a place in future system that we're building. We will maintain as Steve says keep these legacy system going alongside. We will then make a decision in the future about whether those non language courses are going to be turned off or not. And as what I tried to say earlier was Well, for me memorize product is being an unencumbered by this product of the moment that product slows down our development on the main memories product, once we sever that tie in are able to move fast on the main memorize product. Keeping those servers going is not a huge difficulty for us. So it may be that there are things I don't understand yet mean that we do have to turn off those courses. Those courses will not move over to the new members experience because they're not languages. But we don't know we have not made a decision. Those are going to be turned off. So if you heard me saying I'm going to
Thanks for the clarification. What you're saying is you may get rid of those courses at some point but you haven't definitely decided to do so. Even in light of that clarification, I think this question is very important.
Yeah, and so we haven't got a particularly the course content there. We have not made a decision on that. It seems like a really reasonable request. So yeah, it seems like a really reasonable request, and I don't think will be very difficult for us to do. So. I will not make a promise before talking to the technical team. And others that it seems like a really reasonable request and I don't know why we wouldn't do it
then the next question wasn't memorize increase credit to 25 years old, even making a mobile app which you could have made support user good causes as well. Now we actually had mobile apps before that. Were since 2013. It was mainly focusing in on controlling the content or is hard to totally say that but that was the we believe that is the biggest driver and that was the main change.
Then we've got if another company we're trying to create a duplicate of the old memorize community platform, would you cause some legal problems on that leave and we'd leave them to check in with Steve there to see if any eyebrows are raised but no, we live in to we
absolutely we if somebody spun up a new place and said hey, you know we've got a you know, a good a better version of this memorized community. Could you please let people export their content? To that community? We'd say yes. Yeah.
Federal. Keeping in mind educrats keep things respectful. My name is Ben, by the way. I went through that as disrespectful, will memorize commit to reversing course. And start treating community members with actual respect. Yeah, it's good.
I guess I guess I don't know how to answer that. Without knowing how we somehow disrespected anybody. Yeah, so I would want to understand how we disrespected anybody and would apologize immediately and see what I could do to not make that mistake in the future. But since I just want an example,
if you're, if you want to hear example is a pattern of stringing people along and not telling us what's going on, but telling us there have been various times when you've told us what you say is going on and it wasn't the truth. And where you say, you know, you made a strategic decision. You would misrepresent that decision on the forums. I think that was a form of disrespect.
Does that would be Could you give me an example.
I mean, it took two years to figure out that you were deep prioritizing the community courses and why because you spent two years misleading us
so that's that's actually not what Ben said. And not what happened. What happened was, we couldn't figure out the path forward, right. We were sitting there trying to unravel this problem. It's not that there was a decision two years ago that was said, Okay, we're just going to be talking about 20 2014 1516.
Okay, so
we're talking about 2014. My memory was that I said that very straight at the beginning, but if I didn't say it straight enough, my memory was and I said it very straight at the beginning and had a lot of difficult conversations were people who were very angry about that. But if my memory of that is is wrong, I apologize that it certainly was my intent to be completely open that that is we were building our own courses because we needed to take control of the content, because that's how we innovated and I thought that we were really open about that. I do remember that it was very difficult and people were angry about it, but I yeah, it certainly wasn't my intention to misrepresent that.
And then towards that end, we apologize if that was the the net effect. And, you know, it was certainly unintentional. But even without it being intentional, if that was the net effect. We bought guys and we endeavor to try to do better going forward, which is in part why we're here. We're just trying to do better. We're trying to be better.
So today's session is started doing that.
Thank you. Max. Since memorizes, Steve you got to drop.
We're gonna have to go Yeah, so Okay, well, that was great meeting you all, and thanks for having me along.
A bit last man standing. Since members decided to let the user create a course platform with her Would you consider spinning it off to people who want to run such a platform? Yeah. Like if someone wants to take that on? Yeah, definitely. Let's talk. Memorize and memorize have very different goals. Is it possible to do meme world which I generally appreciate? Or even memorize is UGC. I, I think you're gonna see this well, if you take my understanding of what memorizes goals are, as part of this holistic system to learn words and phrases, and then immerse and communicate using those words and phrases. This is all the same goal of learning a language so my mind the first part of this question I don't agree with memorize a meme world or two parts of the same solution. There are a right hand and left hand to fixing this whole problem of how you learn language. So
I'm not quite sure in that basis how to I'm not quite sure how to answer that because I
I'll try I'll try the next one. I appreciate you Ben. Being honest and not promises the exact date of memes going back at least half an approximation. It's a fair, it's fair question. I will try to get you on. I can't give you one now. I will. Talk with the team. It's it is going to be hard right now. Because we don't we don't want I don't I don't want to be promising something that we can't deliver on because it causes problems that I don't get the way I don't yet understand we need to optimize towards moving in the direction we need to go as fast as possible, was making the few fewest mistakes on that route. So that's what we're going to be trying to do. I think it's hard for me to give even an approximation on that except to I will keep you updated. And we can keep yeah we will keep sharing progress. I I think that probably the best I can do.
I understand that memorize. It's about memorizing right. It's not limited to languages. It is but when it comes to language this is about memorizing a lot of words in a short time and effectively remembering them. But wow, man. Well, it's about immersion. As we as you mentioned, it's about communicating,
kind of, in fact, what memorize the main courses where most of our users are and most of our revenue comes from, is about learning words in context. And through videos and audio and through rich multimedia and a connection to people. So it's, it's not spent, they're spaced repetition is not the most important part of it. The most important part is video content and all of that. That's what unlocked that huge wave of growth. And that's that that's the basis of it. The Spaced Repetition is useful to of course,
and that the video part is one way to produce memorize as it's used in the official courses, but it's not possible to to put in videos in in user generated courses. And And considering that a lot of very useful for example, frequency list of words. So these are not about about videos.
If I made the the when you're saying there's there's not possible in UG in user made courses in the system that we are building, that is exactly what is possible in user made courses. Every user made course will have access to all the videos. That's why we're building it that way.
But still, I mean, I mean some some courses a lot are not about videos or immersion or communication, some some quarters about remembering words, just just was so
right and so from the memorize pedagogic perspective, memorizing just a word and its translation is rote but until you can use that in context and see, hear it spoken by multiple people, hear it with different accents create a rich and multifaceted memory of it. You have not learned as much, but in as much as the user the user generated course courses at the moment are useful for teaching you just a word definition mapping and nothing else. That is that is exactly what we are moving away from and we're moving towards a richer sense of what it means to learn a word. And that's been what we've been trying to do from from the very beginning.
I absolutely agree with you that learning a language. Learner language is not about it's not about just learning the words that that's absolutely true. But I think it's important to note to note that many people who don't who don't use memorize solely so they may be having a cause out somewhere else and use madman and to them memorize is a very good supplement. But they're not using memorize solely so they're not expecting to solely use memorize or get fluent but memorize is a very important tool.
Yeah. And so those and for those people, I believe that the product that we are building where they can generate lists of words from their classes, and they automatically they can generate lists of words from films they're watching to help them learn. They can automatically generate those lists of words and get those words learn those words and learn them in context and have suggestions of further films and further podcasts they should listen to, because they know those words, that is a more powerful place to be than the pure user mycoses situation that we have at the moment. So that's why we're building that structure. I just want to go on. Sorry.
I'm sorry. I'm sorry. Yeah, I appreciate that the intention to to to make the ability to for producers to build courses out of for example. For example,
but the message the message I'm telling you, Eunice is that we, in order to build if we try and coexist them, then we can't build the system that I'm talking about the system where you can auto generate your what word lists and get wordless where you get rich multimedia tests, you get tested from context and it's hooked into a full system. We we can't build that system if we optimize for trying to have both. That is what we have tried to do in the past. What we are doing now is we're saying this legacy system where we are going to keep it running in parallel, but we are not promoting it. It has not helped us and it has slowed us down. We are building the new thing. And that's what we're optimizing for. Everything is saying. Building from minority languages, creating resources for them about the ease of making lots of word lists and all of that. That is true within the new system that we're building. But we're not what we are. The tough decision that we are making and that we need to keep on making in order to make rapid progress is to prioritize the new system over the hill. I'm just going to move on to two more questions. I know you're gonna have something else to say. But I do want to make sure that we get I get through everything that just four more here that I just need to get through. So what does this mean for Luma? I don't use memorize. I don't honestly see how flashcard system need to be connected to Luma classes. So Luma is just is an events platform that we have been using for running some test conversation classes that will all be integrated into mem world. Luma is a tool that it's a short term use. Anyway, mem world is where that will go. And use of flashcards is just generating auto generating from the conversations you have in conversation classes. The list of words so the moment has been done manually, and Harry shares those lists back with every one of the lists of words that have come up the but in future there's going to be auto generated into these lists of words in the new memorize product. Would we be able to follow people in the app in future instead of when logging in from our browsers? So interesting. It's not immediately on a robot and following doesn't have a huge amount of use as a feature at the moment. We would certainly love to build in more things that have that kind of social element. And meme world will be a big place for that. But those are going to come well after building the learning experiences themselves. We will lead on creating the learning experience and then kind of social features can can come later but they're definitely not our priority. Will you be including podcasts in the new version? So in men world, yes, there will be a curated podcasts and you're able to see the transcripts while you're listening to the podcast, transcript and translations and the coaches recommend those on a daily basis and your playlists of them. So yeah, that will come. Why UTC have been opponent declined technically in terms of relation since 2016 to 2017. It's actually since 2014 2015. And it's because we made that conscious decision that we were unable to grow fast with the user made courses we could we couldn't make that product work. Maybe someone else can and we're happy to spin it out to them if they want to try, we can make that work. So we move to courses that we created ourselves, and that is what unlocked the growth. So and being able to create new experiences was what allowed us to grow in that way. So that's that lack of quality control and high level content for UGC. I'm not sure what that question is. But, yes, it was the lack of quality control on UGC is difficult. The implementing quality control on lists of words is not far different from just creating the list of words. oneself so it's a it's a difficult it's a difficult process to manage getting quality control good that that is just a fact. But we centralized a single centralized dictionary in each language pair, it becomes a much easier job. So when we used to have centralized wikis in each language,
and I spent two years building wikis, we did, we were able to create a much more effective dictionary for each language pair. Because getting three or four people to work together or making a dictionary works works really well. And you can have quality control on Macs, it's one dictionary that can then feed an infinite number of courses. Whereas when you have to inspect each course individually as you do in the current system, that's a really, that's a heavy lift to do quality control on. So that's another benefit of this new system that we're moving to which is as I say, much closer to the architecture that we had in 2010 2011. That is all the questions there. Because then we get to you're saying when we get to all the questions, do you have more questions that you would like to answer that earnings slide Do
I have an item for the previous question and it also relates to a new question? So I because you said you weren't aware of what counted as disrespect. I was kind of mulling it over and I thought of one of the most significant examples that really dejected me and dispirited me. And so first I'll explain the context. And then I'll say what the disrespect divert was, the context is, you've changed it so that commas no longer separated alternative answers. So that if we had like three words with comments before any one of them could be the correct answer, but then you changed it so that people had to type the whole phrase with comments. And here's the disrespect, though that this broke a huge number of courses because they were no longer usable in the form that they were. But nobody gave us any warning that this change was coming. This change just happened blindly. And a bunch of us reported it. As above, because we honestly believe that was about and one of your developers actually fix the bug and reverse the change. And those courses started working again. And then that developer came back to us and said, Oh, I've been told that this was a deliberate change, and I need to restore it. So I find that that is incredibly disrespectful when you're going to make a change. And even if you have a really good reason for the change, and you're definitely going to do it and it's not and no one can change your mind about it. But if you're going to make a change that's going to break so many people's work. You can't just do it blindly and leave us to think it's above and then tell us after the fact that it was deliberate. You need to give us significance. And the other thing aspect of it is that nobody ever told us whether you even considered the idea of making it up per course setting so that your courses wouldn't have one way and ours wouldn't. Whether you rolled that out for any specific reason. Like nobody ever answered that question.
I can certainly see how that would have felt disrespectful and I thank you for bringing that perspective. I guess I mean, this we're talking. When was this?
Think 2015 20 Year
20 2014 2015. I think it's kind of time. Yeah, and I guess this is not an excuse but to explain the situation internally, or that this is when we were making a lot of those changes. And, and it was it was a really intense time for us. Memories. We had very little money left. Numbers were really flat. We were taking a big risk, which a lot of people were angry about that we were moving away from the user generated courses as the lead. And we weren't honestly sure that we were doing the right thing. But we knew we didn't have long to make these decisions. And we were we were working incredibly long hours. There was a lot to manage. And I guess and I know in fact, we just didn't get to everything and we missed it. And I'm sorry that that was the situation that was on the ground at the time. We just tried to we were trying to move but we weren't we were under a lot of pressure. And we were I think also we were at the time feeling quite kind of scared and defensive. Because we were really close to run at going out of business. And the runway was really short. And people on the forum were shouting at us saying you're stupid, you're making the wrong decision. And we were that kind of like, crap, are we making the wrong decision and of course some people in the team were like, the right you're making the wrong decision and we're like crap, are we making the wrong and it was really scary time. And we in that context, I am, I am certain that we didn't always behave as well as we could have done and I apologies for that. But I will do my best to rectify that in future and make sure that we can and do communicate these things in future.
So here's the related question. If you are detaching the real new memorise from this legacy website, but you're just going to kind of leave in maintenance mode. Does that give you the opportunity to make some small changes? undo some of the things you did that you had to do because the two are linked, this being an example. For example, could you make comments separate alternatives again, on the legacy website, but another thing that you did, for example, was that I have a course that uses images that depends on the course taker only being able to see one image and you changed it so that they can scroll through all images. So there are a number of changes that you made, because the two were linked. And if you are abandoning this, can you undo some of those?
Can we talk in principle, of course, and these are the kinds of questions that we should look at and you're absolutely right, once they're separated. Life becomes easier, kind of part of the problem. We have our memories of the moment is that it's trying to be two products. And they've both got user bases and we're trying to pull them apart. And any change you make is good for one and bad for the other. And it's those trade offs. Soon as we separate it. We're in a much better place. And I mean, let's creatively explore what we can do that like Steve says, We are more than happy to open source that code base like to work out how we can spin that off to someone let them work. We did talk with Quizlet about doing this. It turned out to be a bit complex to do but at the time and didn't quite suit either of our businesses. But we were totally open to that. And yeah, no reason not do that. And I believe you're a developer.
No, I'm an SRE. Try not to be a full time software engineer.
Trying not to be a full time software engineer sounds like you at least know how to code to some degree. But but either way we can. We're definitely open to those conversations. So I can't, I can promise that we are open to new conversations and let's have that conversation. I don't know exactly where we are. Now if I covered all the sureness,
Jonas's question about whether we will be able to organize things in divided things like course levels is a useful one because in order to like curate how I'm teaching a language, making a simple word list is doesn't really convey what the the progression that I'm trying to give someone if you look at my Hebrew course for example,
yeah, we made sure I've got so this is about Will you still have ability to make subdivisions and kind of show a path is so the, the vision that we have for the user made courses in new Memrise? Is that, yes, you'll be able to that there'll be a feature which allows you to not only say, suggest an order of wordless in the terminology we use for those scenarios, scenario, it's just a list of words. I'm going to use scenario to avoid saying list of words a million times. So you'll be able each scenario is just a list of words, you'll be able to the the vision, and this is something not designed up but this is what we have discussed. And what we want to build for this is that you can define a suggested order of scenarios and also be able to pin in with the media that people can watch the podcast that might be good to go alongside it at those points, and so on, in a way, not unlike the old multimedia levels that you're able to do but just a an easier way of putting that together. And combining these things, these features from mem world and these features from memorize. So you're assuming of course,
yeah, I wanted to show you of what I need to set some levels are just list of course yeah, it's not just words.
Yeah. So so the explicit aim within memory memorize mem world ecosystem is that users individuals, so we're in the, again, the terminology that we have there. We have learners and we have coaches and this will be a form of coaching, be able to prescribe a list of scenarios but and then put in between them different bits of multimedia also add comment threads in and discussion friends and so on. So, this kind of thing it won't be exactly this interface. But what you are trying to do here with building up in order, and suggesting how to build up concepts and being able to link to videos, being able to link to podcasts and so on. That is absolutely where we are trying to get to with it. And to Steve's point, we're not going to turn off the hold until we're satisfied with the new. How far along do we need to be before you're satisfied? I mean, that's for us to continue having that discussion. I hope that we can get to a point where you're like, Sure, I'd much rather curate this on meme world using the memorize scenarios. This has got way more functionality than that we used to have a meme on the old memorize. But let Yeah, that is where we are. That is where we are driving to and we just need to move there as fast as we can feel like a pizza.
Its costs right.
Close Yes, close.
Hi, my name is Aleksandra alley for short. And yeah, I'd also be happy to get any input from you for meme world. I'm like the memory architect here. And I have several team members on this call. And so if you wanted to give any insight as to what would be helpful. So you can curate content there or as a language learner. I'm happy to hear it
and also to add to that we this particular feature that we're calling the like our internal name for it is the the notebook that is kind of a learner's notebook of all the different resources that they need or teachers notebook of all the different resources or learning needs that can pull together wordless can pull together other kinds of media, that the moment we're building, we're creating the building blocks that enable that we will see the beginning is choked in the beginning of q2 next year just to work through the designs of exactly what that would look like. As the experience and the functionality that would have and because I'd love to have that conversation with you and get your input on what would be useful and what we needed that Do we have any more questions live I'm
I'm sorry if I missed your response, but I'm not sure if you have to have responded to a question about subdivision because you've said on on the for the forum that there won't be subdivisions on invento so that this has caused some worries that someone has communicated to me privately without all the usual C's, I mean, most of the use cases where we will end because of the inability to create subdivisions and they will be very nasty basically, because of the of the flex of the flat junction that is no longer subdivisions because
I think I've muted that. So what we call scenarios on the new memories are identical to levels on the current one. And so, scenarios will move over as levels. Will you then take those subdivisions, they are still there.
Okay. So how can keep the instructions,
but they so then initially just all of the levels will move over and be available there. Then what were what I was discussing cause there is allowing any user to define a path through scenarios and saying this is a suggested order for doing scenarios. And so as a creator of that you can say this is the right order to do it. And if you're on there as a as a coach which is a content creator within the new Manuel structure, it's someone who create stuff for other people to learn. Doesn't have to be a native speaker. But if you're on there as a coach, then you can define the order of levels that you want people to learn your levels. Equally though someone else will be able to come in and see, oh, I like your level 24. I'm going to put that into my path over here so people will be able to recombine between what are currently causing but there'll be able to pick levels from wherever they want. So so the structure will be different in that sense. So the subdivisions will still be there. And the very first thing that we will build will just allow users on on memorize to create a level or scenario as it's called. Then, as we move on, we will allow them to create a scenario and a suggested path through multiple scenarios. But every scenario that is created or memorize can be used by anyone to go into their into their path.
I can kind of give an example because for example, some some causes are organized in a way that the the users follow the subject there's the levels of the cause designed by the cost creators. So that can be such a structure can can be kept Penikett maintained in the members, right.
So I want to be made sure I'm as clear as possible on this. It is it is the intention is that yes. That cause Creator as a what if they, as a coach on memorize or as they will be able to define a path through the scenarios they create or memorize. So they will be able to reproduce that. However, that won't exist as a course on its own. It will be something that people will have to go and follow that coach and say, What order should I do these in and then be able to follow through like that. So it'll be a slightly different structure, but yes, it will is our intention that functionality will still be possible. It'll be in a slightly different way, but it will be possible.
To add code sounds like that the Creator has to be remain active in Manuel, but as you must know, so some creators have long gone, so they just signed it.
They and so we may actually this is a really good point that I hadn't explicitly considered the if the course creator is no longer active. We'll need to do a bit of work to first of all, bring everything over. And then do one of these learner notebooks for each course to make sure we stitch together the scenarios so that that experience is still owned the same way. And, and that is that is an extra piece of work that we'll need to do to convert stuff over. I don't you have put that up. I will take that back to the team. And I can't give a timeline on it. But now that you've raised it that is clearly we need to do. Thank you for raising it.
So I can I can I can bring your words to the bad USA,
I could have chosen not to I'm going to try and get a transcript of the whole call. And if you're all happy with it, I can post I can post a transcript of the whole call to forums. I should check with everyone who came should send that beginning. I checked that everyone is happy with that, but I'm more than happy for you to pass it on. And yeah, we can share the transcript as well if that's when it's a long transcript for people to go through. I'm happy for you to pass it on but the important the important balance that we want to get there is it is I can promise what our intentions are. But I don't I can't absolutely promise the order that we're going to do this in and I can also promise that we're not going to intentionally break something. Our intention is to bring over the courses into the memory system with we're not going to do that in a way that breaks them because they wouldn't be bringing them over. So we will need to find a solution to to that problem. It's not the very first thing that we're doing, but it is in the plans cool i
If I may I would like to talk a little bit about interface.
If you may, and then I'm afraid I really am going to have to go get my children suffer. Okay, so
another big change in the current better. Besides removing the means is also the revamp of the overall interface. And everywhere I look, I see just how many things the new interface breaks. So first of all, it breaks user written scripts which no longer work in the current better. And it also breaks a lot of features that memorize users itself so you can no longer remove the daily goal in the current user interface. The highlighting of the typing mistakes is no longer working. You cannot press enter to move to the next item during the review session. They also accept of IV i m e typing is broken. The interface of repeat difficult words is don't don't work improperly and so on. Beyond that, it's also much harder to navigate the current dashboard and the current the screen of entering the review sessions. So the question is, which which are the pros of the new interface because everywhere I look I only see the console but why does it have to be created and why does it have to be rolled out in two from so
there's quite a range, which sounds do sound like bug reports and I can check whether they're being worked on but some of those do sound like straight up bugs so needs to be fixed. Some of those design preferences and I know there's been a huge amount of testing on what is easiest for people to understand and what flows do work best for people. And so those are know that change. Changing the site which you're used to using in a particular way is always painful. And I totally understand and accept that. But it isn't an objective fact that something is harder to use, and that the new one new model is harder to use. That is a matter of personal preference. And the testing that the design team have done indicates that this is an easier to use site and serves most of the core use cases better. So I think and yeah, I think that's kind of the best answer I can get. It's it's been designed to serve the users and the users that have gone through beta and the user testing that we do extensively as well as possible and to make the space for all the all of the things that we need to build. And yes, some things it has broken. Some intentionally some unavoidably, like I I don't even know how we would have gone about trying to understand all the user made scripts. It's really I can see it's really annoying that that is broken. And causes earlier point, I can see that I could could feel disrespectful, which it certainly wasn't intended. I believe that the development team are simply not aware of what those scripts are. So maybe we need to think about that. But yeah, I mean, I'm not sure I can do a better answer than that the benefits as yet one thing I can that is one benefit is we totally we needed to rebuild on a news tech stack. So it's an entirely new tax structure and a structure that allows us and is aligned with what we need to build in future. Which to the dashboard was not and didn't help us in that direction. I suspect that's an unsatisfactory answer. But I think that the root of that really comes down to how you're using and how you're used to using the dashboard. And the fact that that doesn't quite mesh with the use case that the new dashboard has been designed for. And I get that that's really annoying. And I hope that we'll be able to make enough changes there. So we'll make it work well enough for you. So I encourage you to give those pieces of feedback directly and if you can, like separate them out. As bug reports. I know Alessio and honor spend a lot of time cataloging those and if you can send them to Alessio in honor we will do our very best to address any ones that we can, but there will inevitably be some that we can't.
And just to say, yeah, we've got the support team. I think everything that you did raise I have already collected and it's in the hands of the of the product team at the moment. So they will look into what's bug reports and as soon as I have an update and I'll click on that thread on the forum.
Okay, but if those bugs aren't fixed in to experience, do you still plan to roll out the new button in this current broken form?
I guess that would be a problem. A question for Jen, who I don't think is here and
has already dropped. I think they're gonna fix as many as they humanly can, and they will be my guess. So, but some of what you've mentioned, there isn't is not bugs and our decisions. And those will not be rectified. But I can certainly take that data question back. And let's honor let's take that question. When I'm trying to answer. Thank you. Okay. So that's a bit of an unsatisfying answer, but I hope we can get back to you with something
if you still have a timer. There's a small point to make, but if I may,
I think that was the sound that the children make when I can remember that she's
okay. Okay.
I tried to be short. So, I noticed that there is a user who created some cause, specific so I over the past few days I I was quite worried about the future not only of my knowledge of courses of minority minority languages, but also causes of let's say, Japanese or French or German, that are about specific topics, for example, who are spawned from a book by a farmer from the novel from France. You know, some some kind of specific vocabularies. So I'm a bit worried worried about the future of these courses because I'm afraid they may not be considered used or popular enough to be migrated or maintained in the in a new platform.
So I feel like we've been around this question a few times. The but I'll just try and reiterate it again. The legacy courses are going to carry on we will carry on supporting those they serve as a not being turned off until everyone is happy with what is happening on the new system, and they don't want those old ones. It's very possible that everyone will be like, here. We don't want this because there's any more because this is a much better place to spend my time. If that happens, we can turn off the old courses. If that doesn't happen and people still want to be learning those causes, then we'll be exploring ways to either just keep it going. Because it's no longer in in the way it's no longer stopping us from achieving our main goals. Then we can just keep going. It won't cost much to run or we can spin it out to another company. We can spin it up to a community we can make a dowel out of it. We can we can we can do whatever we like with it. As long as it's not stopping us from developing the main memory product. We were very free with what's possible. So I think that's the best answer I can give at the moment. We're not we're not about to turn their servers off.
Okay. Good. I'm sorry to say because people are having problem of confidence about after the memes. Crisis, I say because people some people think that memorize making an excuse to think that some features are not useful, but in while in fact they are being used so they're worried because while you're saying that, when causes are used, they are not going to be resolved with hired but that they are saying that oh, they say that they're not used but
do you feel now that you have a sense of what I mean by that like do you have have have I managed to communicate my position on this and
I'm feeling bad. Yes, I'm feeling bad, but I am very sympathetic sympathetic with the people who are having a confidence crisis, the company. I'm sorry to say that that is the case.
Look, I totally understand that. Which is why we're having this call. Because I saw and I've been on foreign phones long enough to know, like, typing answers there was never going to answer people's concerns. Because it is it is nuanced and I can't give certain times I can't make promises. But why I hope to do here is just tell you the story of exactly what we are trying to do and why and hopefully from that context, we can we can get we can just understand each other better. And where arguments come down to what is the definition of USD or not. I mean, it's impossible to know. But you can always argue the data either way, but the important thing is that you know where we're coming from, what we're trying to achieve, and how we'll be thinking about it. And that's what I really want to get us on the same on the same wavelength about
Yes, I'm sorry. Maybe we'll keep repeating myself. But as I mentioned that some causes of I hope you'll remember that those concepts may not be in active usage, but they are very useful.
To ask you to say what you think my answer is about those courses. I'm sorry, what do you what do you think my answer is going to be? too when you asked me to pay attention to minority causes that may not be very used, but are important.
I get the point I get is that do not do not have the intention to retire them unless it's necessary. But um, I I still have some worries about you may think that because of financial reasons, or panic or reasons, that these causes are not not worth the cause that you need data to be maintained.
Maybe this can help. Maintaining keeping the servers on and running is not a huge cost. The huge cost is where that technical structure stops us from making progress. That costs us an enormous amount in opportunity cost and in wasted work. Once we've separated that out, the cost of continuing to run the servers and fix occasional bugs as technology changes is really trivial. So that's just that's trivial. So the only situation where I don't know when you say things like the only situation reality as a way of playing with that, but if you it seems incredibly likely to me that once we separate the two if we can make progress on the main memories and Manuel product, I don't immediately see a reason why we would turn off the servers for the Legacy product, because it's not going to be very expensive. And in the context of the overall budget, the cost of maintaining that is really not very much. So I don't think that'll become a problem. The problems become the problems come in. Because right now that product is stopping us from developing for making progress on the main product. And what we found was when we stopped making progress, we stopped. We stopped shipping new stuff. We stopped increasing user numbers so fast and we stopped having the money to then reinvest back into making more features. And yeah, I really do have to go now.