Episode Transcript
[00:00:03] Speaker A: Hello and welcome to Localization Today, where we explore how language, technology, and community converge to unlock ideas for everyone everywhere. My name is Eddie Arrieta and I'll be your host today.
In this episode, we're diving into how AI powered machine translation has reshaped professional translation workflows since 2012, and the story behind Global east by MEMOQ, winner of the 2025 COTI Award in the machine translation category. We'll look at how Globalist came to be, how the MT landscape is evolving, and what this recognition means for memoq and the broader industry. Today, we are joined by Agnes Varga, CTO at memoq, who has spent more than a decade developing and shaping the platform, from software engineering to product leadership. And Gabor.
[00:00:56] Speaker A: Senior Product Manager at memoq, responsible for MT and AI. Gabor's career spans three decades at the intersection of language and technology, from translating SAP software to co founding Globalis and leading its integration into memoq. Agnes Gabor, welcome and thank you for joining us. Gabor. I hope I did all right with your last name.
[00:01:25] Speaker B: That was perfect. Thank you.
[00:01:26] Speaker A: Thank you very much. Agnes, how are you today?
[00:01:28] Speaker C: Thank you. Thank you for, for having us here. And welcome Eddie. Or hello, Eddie, I would say. And. And welcome, everyone.
I'm pretty fine. I'm excited about this podcast and I'm glad that we are here together with Gab, where I think it's also very exciting.
[00:01:44] Speaker A: Yes, yes, yes, yes, yes. Very exciting. And I want to make sure we do justice to. To the conversation so that those that are coming new to the audience, coming new to the conversation, can get a sense of who Agnes is, who Gabor is.
[00:02:01] Speaker A: How would you explain to someone who doesn't know what it is.
What it is. And Agnes, maybe you can help us understand. How would you explain what you do at memoq that connects with global ease and this whole conversation? Maybe, Gabor, we can start with you and we go with Agnes. That's okay.
[00:02:19] Speaker C: Okay, yeah, thank you.
[00:02:21] Speaker B: That's absolutely fine. So, yeah, let's start with my. Myself. So I would say, yeah, I'm a kind of localization industry veteran. It's a little bit sad to say that because time is passing so fast. So I just recognize that I'm over 30 years in the industry.
I played pretty much every role you can have. So I was in the. As a translator, then also the. The proofreading IT manager at some large, large enterprise, leader of the localization team responsible for Hungarian localization and then managing my own company, both localization and, and then developing and.
And so basically I, I think I, I had many goals and I was very lucky. I feel very lucky because basically I grew up with the industry because at the beginning, in the Beginning of the 90s, software localization was basically absolutely new. So before it, this simply didn't exist.
So, you know, when software was shipped out without any language, so it was a separate product shipping out a language variant of a product for shipping out a different product on a different floppy disk. So let me show it about myself.
[00:03:42] Speaker C: Good to hear these things about you, Gabor, because some of this I didn't even know. Even though with that we know each other.
[00:03:50] Speaker A: That's really good. Agnes, if you could allow us, for those that might not know what, There are many CTOs and depending on which organization they are at, you know, they developed different, different, different elements.
How would you describe what you do at memoq?
[00:04:07] Speaker C: Okay, so my job is.
[00:04:11] Speaker C: I told my manager today that CTO for me doesn't describe what I'm doing because I don't only do technical things, but I also oversee the product. So I also oversee the product roadmap and the technology together at memoq and all our products. So I lead this team we call production team, where we actually produce our products, which is memoq and Globaliz and everything else that we are coming up with. And so the whole development team and the product team, Mercabor is, and the QA team and design team, they all belong to my big group. And that's what I lead at the moment. And I started out as a developer at memoq. I came here. It's a very interesting thing. Now that Gabor was talking about his career. I don't come from the localization industry.
I was a lecturer at the university and I wrote my PhD about the accessibility, the acceptability of machine translation. But already more than 10 years ago, and I thought that topic was already kind of obsolete and nobody cared. But somehow now it is coming back again, this topic. But my PhD is pretty old, so it is obsolete. But still the topic is kind of interesting. And this is how I came to memoq through this PhD topic actually. And then from developer, I became product manager like I were, and then cto.
[00:05:36] Speaker A: It's really interesting to see how, you know, lives and careers develop. And that is really good. Today we're talking as well about the CODI Award.
Can you explain what the COTI Award is?
What's the relevance of the COTI Award and how did it come about to get this recognition?
[00:05:56] Speaker C: Can I just say a few sentences and then I give the floor to Gabor because he's the main person behind this whole thing that I just wanted to say how proud I am of the globalist team that they managed to win award which is about the software. And I think it's a prestige and I'm really glad that they got this also as a recognition for their past achievements. So not only the achievement with memoq but already for their, for their achievement as a, as a separate company. So I just wanted to add this that I'm really proud of the team and Gabor. Now I will let you talk more about this award because it's your award.
[00:06:34] Speaker B: Yeah. Thank you. It's r. It's the, it's. It's. It's a camera. To join the Ford. Of course. Yeah. The Kodia V, you can look at it like it's, it's a kind of Oscar prize for, for technology. So that's in short, that's basically a recognition of, of. Of some, some achievements you have. And I'm really proud about it that we, we did it. It was for me very, I mean it, I was, it was, it was. We just. Just.
I mean I was just informed that we nominated globally. I mean the company nominated globally. It was. So the first step was denomination.
[00:07:10] Speaker B: We had some very tough and intensive.
[00:07:15] Speaker B: Rehearsal with some experts and yeah. And then the result, the outcome was that we were one of the winners and yeah, that's exciting. And of course the team was very proud about it as well. So. And then finally it's there they're price their recognition of their work because the actual developers or the actual people behind it, they are making the products that is really good.
[00:07:42] Speaker A: And what does it mean in terms of the evolution of the product that you are working on? What does this recognition signify? What is being recognized?
[00:07:53] Speaker B: I would say this, I mean specifically what during.
[00:07:59] Speaker B: When we, when I presented and when we had the discussion. So what the product can do is.
I think one of the reasons why.
[00:08:12] Speaker B: We were in this, we got in this position was the.
[00:08:18] Speaker B: Very unique approach about custom what we are calling dynamic prompting. So I don't want to go into too much into the technical details but nowadays basically with the large language models.
[00:08:32] Speaker B: Taking over the main rule in machine translation, the race is basically about the best prompting. So this is my. In short, that would be the summary and we have a unique approach. We are combining different technologies. We are still utilizing neural MT technology for extracting the terminology. This is one of the main points I think which is quite unique that this helps the engine to provide the Right content, right context, the right terms, even if they are components.
[00:09:07] Speaker B: And so neural machine translation technology is combined and used to retrieve the custom terminology. Then we have custom prompts, what can influence the output. We have the reference translation.
[00:09:21] Speaker B: And keyword list and all that together is basically providing the final translation. And this approach, I think this I believe rather unique approach was in my, as I believe one of the reasons why we won the prize.
[00:09:39] Speaker A: That is really great and kudos, Gabor, because it sounds like being at the very edge of innovation and trying really complex things and really putting systems together. And it brings me to the conversation which we'll dig a little deeper in a bit, which is ecosystem.
And it seems like, Agnes, if I'm allowed to say this Kodia word is also a great signal about the adaptation of global ease with memoq. And when we were at the memoq fest, we were talking about the transition and, and the projects and the plans. Now there is an award. Agnes, could you give us a few words about, you know, what it has meant for the memoq team to come in, you know, this partnership with globalis?
How did MemoQ integrate into the broader ecosystem? How did MemoQ integrate globalism to this broader ecosystem to get these as a result? Right. Which is like great innovation.
[00:10:49] Speaker C: Okay, so I think I start from the point where memoq, few years ago or until a few years ago, we had this position that we didn't really want to go into other fields than the main TMS technology. So we rather used external MT providers with memoq and Globalist was one of these external MT providers instead of creating our own machine translation system. And then LLMs suddenly burst into this picture where we saw a lot of fantasy that this may actually be very useful for our users and we should go in that direction. And that's where we created agt, the, you know, the RAG based translation functionality. And then there somehow seemed that there is a kind of a gap that we cannot really provide any machine translation. And then we also needed some knowledge about this whole area and there we had a good partnership already with Globaliz and Gabor and we always enjoyed this whole partnership and then the connection between us. And we thought that this would mean a great synergy between the two companies where we had a kind of a gap or a missing hole in the whole ecosystem. What you were saying that we didn't have a machine translation service at all our own. And then here, this is what we wanted to feel and could also feel. So we have now neural machine translation. We have neural machine translation combined with LLM technology and also we have LLM technology. So I think this is a very good combination of all the options and also gives a lot of flexibility for our customers because at memoq I think maybe I will repeat this a few times, but our whole core thought and core, what we live by is giving value to the customers. So to give something that is useful and that increases their productivity, that they can use and all layers every participant of the ecosystem.
And then the Globalist team came and joined us and I think this was integration in two ways of integrating the products into the portfolio.
So now we can actually provide a machine translation service that we could not do before. And also the team, I think the team was a great addition to our development team and Gabor and his team joined memoq as well and they were a great addition and they brought in huge expertise in this field in the field of machine translation, in the field of LLM technology and research. And we do a lot more research and development now than we used to do before.
And also we managed to increase the team of the Globalist team where Gabor is the PM and we have more developers working in that team and also we're working on extra very exciting products that will appear next year.
[00:13:49] Speaker A: Yeah, I don't know Gabor, if you want to say anything. Yeah. Just go for it.
[00:13:52] Speaker B: Yeah. If I may comment on that or add my thoughts to it. Yeah, thank you. I would say that it's also a very good. If you speak about the ecosystem and how industry is evolving. I think this is so this, that we joined memoq, that the Globalist team joined memoq is also a kind of indicator for these changes we have in the industry because that's the TM or machine translation is now getting in the central part of the processes and this is basically kind of indicator for it. And, but this is just the beginning. So the.
[00:14:33] Speaker B: The, so the empty. So the transition from the translation heavy processes were empty. Just an additional translation hit will switch to completely machine translation heavy and centralized translation and translators will mostly do post editing and quality assurance before and after the project.
So that's the, that's the kind of, of impact I see on the ecosystem.
And from, if you, if you look at the, from the buyer's perspective, I would say that the, the most interesting to me is to see how the buyers will handle this situation that they can get now really cheap and really great results from machines. But still I, I'm, I'm, I'm interested if they will recognize that they will still need the humans in the room loop because this is one of the debates we have in the industry. I think that. So someone said, okay, we don't need the humans in the loop anymore or that can be automated fully. I personally believe that to some extent it might be true that, but still there is a lot of room for humans and this is exactly what is leading back. So this is giving the TMS system still an important role. So TMS systems will still play a very important role in the future in the overall processes, in the overall ecosystem for the users or I mean from the other side of the story, the translators or the LSPs, they have to adapt, they have to recognize that they have to offer new services. And exactly this focusing on the quality assurance parts, terminology work, etc.
But still in the center will be somehow I think.
[00:16:22] Speaker B: This tools extended.
But maybe this is some other part we can go into that detail.
[00:16:32] Speaker B: So there are some huge significant changes in the overall workflow and the usage of the systems for it. I'm just at the keyword I mentioning Agent dki.
[00:16:44] Speaker B: So that's in short, and this is as I mentioned, this was exactly a natural process to us and this is why I'm happy that we are now part of the memoq team because we are now basically fit this very natural process which is indeed happening in the industry.
[00:17:02] Speaker A: That is really great to hear, Gabor and Agnes.
And of course I'd say in this conversation the elephant in the room is, and I'll preempt that by saying that there is a huge evolution in how the conversation around AI has taken place.
There is a huge correction on the expectations.
And I think there is also, in that correction of the expectations, much more innovation that was originally expected as well, much more creativity and diversity in execution than it was expected to begin with. So one of the boldest claims at the very beginning that continues to be in some less sophisticated places is the TMS is dead, the MTs are dead and human in the loops are no longer needed in the conversation of the ecosystem. That elephant in the room conversation is empty. That and I'd say the more abstract question to me is what role is MEMOQ and Global east playing in the overall technology ecosystem that allows you to be confident that this approach and this observation of how MTS evolve and the evolution of the needs of the customer, like Agnes was mentioning how all of that, it's how you are positioning yourself in all of that. And please take it as you might. I know I said a lot of words, so take it as you might and give us your honest opinion as much as possible.
[00:18:38] Speaker C: Should I start? I will start. I think yeah.
[00:18:43] Speaker C: This is a very difficult topic and I think it is difficult in this industry especially because somehow this industry is very, very much affected.
I think it's more like a wishful thinking of top level.
[00:19:00] Speaker C: Companies who would want to get the human out of the loop, who would get everything out of the loop and do everything very cost efficiently.
But I don't think this extreme is the truth. I think the truth is somewhere in between and we do need the human. And it's also turning out that not everything is as bright as it seemed at the beginning. As you said, there's massive expectation management.
Not everything is that bright. And I really in many, many scenarios this.
[00:19:36] Speaker C: Merged flexible solution of NMT and LLMs work really well in the very specific scenarios. So there are many scenarios where this technology works better and there the other technology works better and also there comes this question of access.
I may always say accessibility. It's not. It's acceptability. Acceptability where how much are we willing to accept bad quality and how harmful that is? I think this will also be. It will balance itself out probably eventually. So it will take take a little bit of time where the excitement dies away a little bit and then we get back to reality and then realize that what we actually need and what we don't need and how much cost can we save and how much we lose with that saving.
So things will balance itself out, but I think it takes a bit of time. So now everyone wants to solve everything with agents and AI.
But.
[00:20:39] Speaker C: I hope or I think it's not possible in the end, but we will see.
[00:20:43] Speaker A: Thank you, Agnes Gabor.
[00:20:45] Speaker B: I think, yeah.
In connection to that one of the major changes we are already, what we can already see is that this approach of risk assessment, so two categories because before we had translation was translation and the quality was always human quality. So there was no debate about differentiation in quality. But nowadays it's a very typical approach.
[00:21:15] Speaker B: That'S content is categorized and depending on the risk category you can have different quality and different workflows and different tools, different processes behind it. So if you have a high risk content then of course it's undergoing a full post editing workflow and with probably different tools. But you have content you might have contented. That's a very big portion actually fortunately or unfortunately depending on from which perspective you are looking at it. For the translators maybe it's an unfortunate. But for tool providers, I mean especially for empty providers, it's of course something Good, but so a significant portion is moving from the mid to low risk category where human post editing is not necessarily required because it's simply not a big issue. If there is some mistranslation or even wrong contents may not harm anybody and that for that scenario, then the automated workflows are still fine. And this is something which is now getting more and more I think practice so to have these different categories. But there will be always the room and there is always portion of the, of the, of the contents to be translated, which is, which is falling in this, in the, in the high risk and the human in the loop and.
[00:22:45] Speaker B: Really high quality category. So from that perspective, especially if translators are. And LSPs are specialized in some domains, then I think they are still.
[00:23:00] Speaker B: They will still have some work to do in the future as well.
[00:23:04] Speaker A: Absolutely. And I think from your answers, Agnes and Gabor, I kind of can infer where you think people should be looking into.
[00:23:16] Speaker A: I think we are all in the industry realizing that we're coming back to the conversation of yes, human in the loop. And then someone in that event ask yeah, but where exactly in the loop.
[00:23:29] Speaker A: And way before the loop, right in the loop. Way after the loop, Right after the loop, we're in the loop.
[00:23:37] Speaker A: And I think it's a great question and I think it's a great conversation to have around the developments that we're going to see in our industry of what's valuable and what's not as valuable.
And we have this whole conversation about training data. And then you have synthetic data, and then is synthetic data as valuable as real human data? Probably the answer is no. To what degree? In which context? And I think we are evolving in so many different ways. You are right at the center of those evolutions for the two of you. If you were to talk about those developments in whether technical or general economic, for globalization, content translation, linguistic teams, data teams, what would you say they should be looking into? Where should they have their curiosity going into?
[00:24:29] Speaker C: Go ahead, Gabor.
[00:24:33] Speaker B: It looks like you.
[00:24:33] Speaker A: Want to do it.
[00:24:34] Speaker B: I would say that if in a short term, I mean, sorry, in one firm, it's data curation. So in my opinion this is one of the hottest topics. Data management, data curation and especially of course linguistic data creation. Because I just had some discussion so that the.
You, you, you, you should never Forget that the second L in LLMs is representing language. So language will still be very important.
And, and, and even, even if the. So you. And maintenance, maintenance, maintenance is required. So we know all that in the. It's In. In the. In the.
In every aspect of the. Of. Of of life. The maintenance. Without maintenance nothing is running. So. So. So are not LLMs so you can have perfect translations. But over the years, over the time language is changing, terminology is changing, so style is changing. So there you have to somehow catch it up and you cannot do it purely automated. So.
So this is why you need data curation and data. Data data management.
If somebody is. And I think the winners will. Those will be those who are recognizing that and who will change. I mean I know more talking about the vendor, sorry about the buyers. So the translation bias because it's finally up to them recognizing this that it's not enough to start translation and running everything on AI let's say, let's talk about AI.
[00:26:07] Speaker B: But on the midterm and long term they will need.
[00:26:14] Speaker B: Curated data, qualified data and not only so linguistic data in a very wide meaning. So not only written content, but audio content, video content. So any, any. Any kind of. Any kind of format and any kind of any from any aspects. So for example labeling proper labeled content, et cetera. So this is, this is, this is.
[00:26:38] Speaker B: What'S coming up that means that to me that the translation will change in that extent that it's not real. It's not only about.
[00:26:49] Speaker B: Providing the right translation. So it's. It's the translation work as it is will. Will still be there but it will be the less significant. But this data creation work that will. That will be a very distributed future and both for the buyers but also for the vendors. I mean the service providers and the translators, this is the fields they can then jump in. And of course from tool perspective this means that the tools, I mean the TMS systems the focus will switch more to that. And in our example is for example is this what we have the in country review tool.
[00:27:32] Speaker B: Which is basically a very good indicator of this changing process is that now you not necessarily need translators, professional translators in the loop. Maybe you need an expert who's sitting in the country at your subsidiary who can check the text. And they are not necessarily professional translators, but they are domain experts and they will do some part of this work. And this is a very good example I think for that.
[00:27:58] Speaker C: Yeah, I can just add that I think also if you are in somehow language expert or in this industry, your interest may be creating something or then you will not probably be a very good data curator. So I think you need to also look at your interest and maybe where another option is exactly the automation part and the process is to looking at Processes because I find that more creative than actually data curation.
So I think creating a new process or a new automation or looking at how we can do something more efficiently or effectively, I think that is somehow also creating something.
So that's, I think, also an area.
And another thing which I think we tend to forget about is that we still want humans to review things. I can also bring the analogy in software development where the trend is now to generate code with AI and then just check the code and review the code. But we tend to forget that to actually review the code, you have to also be able to write the code and to understand what it is. And if we don't train our people and we somehow lose this expertise of training junior people to being seniors, then we will also lose the capability of reviewing. And I think it's the same for translation. So if you don't know what translation is or if you don't know languages, because in a few decades, or I don't know where we will be in few decades, though let's not talk about that. But in a few years, we keep losing all this expertise, we might be in trouble in the end and then we will have no choice but to accept the quality that we get or, and the risks that we get. So we also have to maintain these experts expertise somehow. But that of course costs money and who wants to pay money? But anyway, that's another question for companies.
[00:30:01] Speaker A: Yes, Agnes and Gabor, thank you so much because this is really insightful and in a way it sounds, oh my gosh, it's so obvious. We assume everyone's gonna do. It's like, actually no, we will go without it without needing to have data curation until it's painfully.
[00:30:22] Speaker A: Say.
[00:30:25] Speaker A: When we start seeing that it's not bringing the revenue result at the moment when we are like, what is going on? It's like, oh, your content is so plain and so like everything else that you need to add textures. Like, what do you mean by adding textures? Like, well, there are these linguistics and culturalization professionals.
Not everyone will be able to afford that.
Not everyone will be able to add that level of texture. Is this why you're so bullish on empties and customized empties? I don't know. Gabor, Agonist, what is a costume empty?
And how does it differentiate from something generic? And I think that an answer to that can help us understand all these other opportunities and where to bring humans in the loop and whatnot.
[00:31:12] Speaker B: Yeah, thank you. Yeah, that's a good question. Because it's indeed There are different scenarios and I'm always saying that quality is not absolute. So it's not nothing that's there's no the best engine in the world and the best solution or technical solution or the best LLM or what. So it's always depending on what is your scenario.
Customization custom engine. So customization options are important.
[00:31:43] Speaker B: In very specific scenarios.
[00:31:46] Speaker B: Please.
[00:31:49] Speaker B: Give you some examples. For example, there's a use case we have to me it's very interesting. It's a large public transportation company.
[00:31:59] Speaker B: And they have a very specific content because they are using MT for.
[00:32:06] Speaker B: Automated translation of public announcement. For example, if there is some of the tram lines are there is an outage or something and.
[00:32:16] Speaker B: The people have to switch the train or something like that then it's announced in an app and this will be translated automatically.
And this sounds very generic but it's not absolutely not generic content because you have many specific for example station names, station names sometimes must be translated. Station names sometimes shouldn't be translated.
So it's not absolutely not trivial to have some. And this is I think one of the good examples for custom engines where you have, where you can add custom data, training data, custom keyword list where you can fine tune the results of the difference until you can get really the desired results.
Or another example could be technical documentations e.g. linux manuals.
[00:33:08] Speaker B: For Linux user menus. You have very mix of comments, you have very heavy formattings. So if you have a constant content with a very specific.
[00:33:21] Speaker B: Strict this is important strict terminology.
This is again a very good use case for custom engines. The border between generic and custom engines is now merging. In the past it was more significant nowadays with the LLMs and the. So customization options are getting better and better. I mean more efficient with the LLMs now you have much more and easier options to do that.
[00:33:47] Speaker B: But still, still a generic LLM. Even if you are providing some terminology risk to it it's not necessarily always providing the desired resources result. So, so this is basically this is. This is. These are basically the exam based and this is the direction we are working on looking looking at to to find the right options the right balance between easy customization and and, and. And. And. And. And. And and and exact results. Especially if you just wanted a challenge I can mention compounds. So for example, if you have the expression service provider and you want the expression service provider translated in a specific way but you have the other expression language service provider and language service provider should be translated again in a different way. So this is for example a challenge where you probably need a custom solution custom engine with a specific customization and solution.
[00:34:48] Speaker A: Thank you Gabor. Agnes, I don't know if you have anything to add, but if we come back to the recognition, the Cody Award and the conversation we've had looking into 2026, what's coming, what are you excited about?
[00:35:05] Speaker C: Actually we are excited about a lot of things. I think MemoQ as a company and the product is evolving.
[00:35:14] Speaker C: And I'm really happy to see that and that we are coming up with a new product next year offering generic translation for a more general audience that is called Fluent that will come out next year and also we will concentrate on text transformation with the help of AI as well and also more flexible workflows and we are really, really excited to offer this to our customers. And of course we are still working on our cloud offering which is an absolute must at the moment. But it's really nice and I'm really glad to see how it is improving and what we can give to our customers about the world.
Well, let's see how things will turn out and how the whole LLM world and AI providers will work in the future regarding offering their services and how the whole pricing models will change.
I'm not sure I'm excited to see all this. We will see if it's exciting or if it's more difficult because I think things will probably change from the current situation with all the bigger LLM providers or AI generative AI providers as we can see it in the news constantly.
[00:36:36] Speaker A: I look forward to seeing how this, Yes, I look forward to seeing how these things evolve.
Thank you so much both of you for your time. Before we go, is there anything you like to add for our great audience?
[00:36:49] Speaker C: I think I've said what I wanted to say. Thank you. Thank you for listening.
[00:36:53] Speaker B: Yeah, maybe our final thoughts, at least from my perspective, what will be the next big thing and I'm very excited to see what this will bring to us. I know this is nowadays a buzzword, but AgentIQ AI will definitely be the second evolution I believe and I'm very excited to see what just as an example, this will as I believe completely change the user end user experience, how we can deal with use and also in the translation industry because the UI can switch completely to instead of clicking on UI elements you can simply have a dialogue with the system which is a completely different experience. And this is one of the directions some of you are also looking at. Of course this is something on the long term, but for me this is one of the from the technology point of view the next week. Singh.
[00:37:46] Speaker A: All right, all right, all right. Thank you. Thank you, both of you for joining us today here at Localization Today.
[00:37:52] Speaker B: Thank you very much.
[00:37:53] Speaker C: Thank you very much.
[00:37:58] Speaker A: All right. So thank you, all of you, for listening to Localization Today. Once again, a special thank you to Aknis Varga and Gabor Cheney for joining us and sharing their perspective on the evolution of a AI powered, the development of Global Ease by memoq and what the Cody Award means for the future of translation technology. To Learn more about MemoQ and Global Ease by MemoQ, visit memoq.com that's M E M O Q.com Catch new episodes of Localization today on Spotify, Apple Podcasts and YouTube. I'm Eddy Arrieta with Multilingual Media. Thanks for joining us and we'll see you next time. Goodbye.
