The Bicultural Podcast
The Bicultural Podcast celebrates culture and identity, exploring cultural differences and shared experiences to inspire stronger business relationships. The Bicultural Podcast is published biweekly and is hosted by Janina Neumann, a bicultural creative and business owner of Janina Neumann Design. Get in touch with Janina: hello@janinaneumanndesign.co.uk
The Bicultural Podcast
What AI does to language and culture with Sabine Harnau and Holger Schueler
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
In the second episode of season 6, Janina Neumann interviews Sabine Harnau and Holger Schueler, Managing Directors at From Scratch Communications.
This episode will give insight into
- Why AI's default voice is male, American and Protestant
- How the Bible became the training data for hundreds of under-resourced languages
- What happens to your organisation's culture when nobody has documented it
- Why shadow AI can become a culture problem (not just a data protection one)
- How AI-generated copy can hide internal disagreement instead of revealing it
If you enjoyed this episode, please review, subscribe, and share with others. Your support means a lot!
Get in touch with Janina from The Bicultural Podcast: hello@janinaneumanndesign.co.uk
www.janinaneumanndesign.co.uk
Welcome to The Bicultual Podcast. I'm delighted to be joined by Sabine Hanau and Holger Schueler, managing directors at From Scratch Communications. Hello both. Sabine, would you like to introduce yourself? Sure.
Speaker 2Thanks for inviting us to be here today. Yeah, I'm Sabine. I started from scratch in 2017 in the UK actually, and moved it to the Netherlands in 2018. I think Holger joined in 2022, and we started out mainly copywriting and now have grown to cover much more websites and culture strategy for expert-led brands. So the more nerdy, the better in our books. We are also certified geeks ourselves. And yeah, so really, really looking forward to our chat today.
SpeakerYeah. Thanks, Janine. I'm super excited to be here. Thanks for the invite. I'm Holger, Holger Schueler. Joined from scratch in 2022, as Sabine said. After an illustrious and uh diverse career in software as a service companies uh for more than 16 years, um, from all the sort of uh weird things of the first steps of video conferencing to live video streaming and in the end uh digital design collaboration and design system platforms. Um so uh yeah, after those mad 16 years uh working with a lot of interesting brands globally, um, I decided it was uh time for a little change and uh continue to um help organizations remove silos and work better together and strengthen their culture, and uh that's what got us together in terms of uh ticking from scratch in that direction, and uh that's what we're here for today.
Speaker 1Lovely, thank you, Sabine. How did you come up with the name from scratch? Goodness.
Speaker 2Actually, um, I came up with the name before I knew I wanted to do my like start my own business. Um at the time I was a member of 26, which is uh an organization in the UK that's all about uh fostering a love of language and writing. Um and they did a workshop about marketing yourself as a writer. I was working as a copywriter in an agency at the time, and I wanted to do more to build my own personal profile. And so I attended that, and then there was a whole uh exercise around finding the common thing that connects all the different threads of your personality, hobbies, ways of working, and so on. And for me, the connective tissue was really the from scratchness, like restarting my own career several times from scratch, moving to a different country, starting from scratch there, um, building things, but making things from scratch, whether it's cooking or uh crafting, and kind of going back to first principles is something that I really enjoy to ground myself in the actual philosophy, but also knowledge and science that we have accumulated as people, humans, humanity, but also because I think that um it's a really clarifying factor when uh organizations and people get confused, it's like it's a good anchor to go back there and then build up from there again. So yeah, uh it all came out of that workshop, and somebody, my um my mentor, Heather Acheson, uh, then said that would make a great name for a copywriting agency. And I was like, nine, six months later, I was like, okay, I think I've got to try this. So yeah, you never know. Any day can change the trajectory of your life completely.
SpeakerI still remember seeing that name for the first time. Um, I was like, yeah, that's it, no doubt. Like it it's gotta be this. Um, it it still leads to interesting situations where people don't get it and uh they just think we're called scratch. Um, and when they get an email, they're like, Yeah, the email is from scratch. So but uh yeah, I think a majority of people get it. And um, it's funny how often it then shows up even in first conversations, and people mention it uh unintentionally without actually thinking of us, but more in terms of where they want to take things and how they want things to be approached. So it's uh it it seems to have lasted um the almost 10 years, and I think it will take us hopefully into the next 10 years at least.
Speaker 1Yeah, if it's part of natural um communication, I think that's always really positive. People get the idea without having to explain it. I mean, yeah.
Speaker 2And I love being in a call where we're talking about potential projects, and then people say, you know, in a regular sentence, just uh the phrase from scratch, and then their face lits up, and it's like, whoops, I just said it, you know, that's what we're talking about. So yeah, it's um uh it's definitely an anti-blueprint kind of stance that we also convey with this. And I think uh in keeping with the topic today, talking about AI and culture, I think that's something that is increasingly popular with the kind of people that approach us for work, you know, getting away from the kind of prefab um approaches and ready-made answers because they very rarely fit a real messy organization.
Speaker 1Yes. And I'm particularly interested as well, Sabine. Where do you think you've uh gained the skills from, you know, starting something new? Because most people would really struggle with that. And I can imagine that you give um your clients confidence in actually exploring other options through the process.
Speaker 2Uh, I wish I had a tidy answer to that. Um, I'm the child of two civil servants, uh, so I've definitely not learned it from them. Maybe my grandmother, who loved the kind of move house, um, and I think she lived in five different houses when I was a kid, even. So maybe it's something almost genetic or passed down from her. But I think it's partly also my wiring. I get impatient and bored very quickly. If like I can't do just you know, road work or uh routine stuff. And my original career was in teaching, which is a good mix of having a stable timetable and you know, a year that's kind of organized. But then within that, there's so much freedom, and every uh student, every pupil, every parent, every other teacher is completely different. Every year it brings new stuff that you can teach and new challenges. So I think that's still kind of the um the ideal balance that I'm looking out for. Um, but yeah, I have to say, like, I think it's partly just because I love to a certain degree, I love adventure and I love the kind of the excitement of the blank page, um, the the empty horizon and being able to shape something that's unimpeded by unnecessary, like we've always done it this way, kind of thinking. Lovely.
Speaker 1Oh, that's so interesting.
Speaker 2And Holger, I have to say, like I would be I would miss the obvious thing because Holger's also very much like that. So I think Holger brought that out in me. Um, maybe the the seed, the germ was always there, but then with Holger's amplification, um I've certainly had more courage to try things uh than on my own.
SpeakerI think that's an interesting one because I initially certainly I think almost 10 years ago when from scratch started, I didn't even realize how much of that was my standard operating procedure. I think it takes time and some self-reflection to realize um our wiring in that fundamental way. But yes, um even my my career prior was always about disrupting something that was the standard for many decades. I mean, video conferencing completely changed the way we worked, and that was even before COVID. COVID completely made it a new default. But um when when I started uh in in that industry in in 2006, um hardly anyone was was doing video conferencing because it was really expensive to do. It required specialist equipment in some boardrooms, and that was it. Um and now people do it with their loved ones from their phones while they're in the supermarket. So um that completely changed the way people worked. Um then even the next chapter of video streaming was like it was television, and it was like programmed pre-programmed television on five, six channels that then turned into like satellite TV, but internet video completely changed that industry upside down, and now that's the status quo with Netflix and Apple TV and YouTube. Basically, nobody like we still sell TV sets, but they're not connected to an antenna or a satellite dish, they're just displays for our digital devices. Um, and I guess then in the last chapter of my SAS world, and I don't want to sort of talk about that for ages because what we do now is is obviously the thing that drives me, but that was also like how can we actually create that sort of reinvention and innovation at scale? How can we make that not the result of some random acts of fortune, but how can we create processes that reliably create innovation and new things and change ways? Basically, change is all there is, um, which is I think an important realization. Like the human nature seems to urge or seems to require consistency, but actually at the same time, we can't really live without constant change and evolution. It's it's why we're still around and hopefully will be for uh millennia. Um so uh yeah, that's all what's in the name. That and then some.
Speaker 1Really interesting. And I guess the latest disruptor is AI. So I was wondering, you know, what parts of human culture do you think show up most clearly in the way AI communicates?
Speaker 2Well, it kind of comes with the fact that the AI we're talking about is most likely LLMs, large language models, and that already has the word language in the name. And really what I studied at uni was cultural studies and linguistics plus educational psychology. So from that point of view, I think we need to when we when we talk about um the fact that AI learns language from humans, it doesn't learn it the way a child would during language acquisition, where they try themselves out and they try to copy what the grown-ups are doing. Instead, the training, of course, goes uh through statistical likelihoods, right? So it's about producing words that are likely to appear next, and that relies on huge amounts of uh training data, and that is US-centric. I mean, there's a lot of uh research that points at the early ways that AI was mainly trained. Unfortunately, the later models are not very forthcoming about where their training data is coming from. But there's this paper um from 2022 uh with the very fitting title, The Ghost in the Machine Has an American accent. Um, it's pre-print only, like most AI criticism. Uh it's not peer-reviewed and printed yet. So there's a little bit of a disclaimer there. But you can say when you look at what AI produces, especially if you use it in English, the model's default is American professional class English. Um, and a lot of the training data that we do know has been used is Common Crawl. That's like a small nonprofit that's been downloading a copy of the public internet every month since 2008 and giving it away for free. And almost every AI model that you've ever heard of was built on that. It's the backbone of most LLMs, at least around the time that GPT uh launched. And we all think, like everybody who looks into this thinks it's still the case. Um around 45% of that common crawl is English. Um, as of October last year, English was used by almost half of all websites on the internet. The web the internet is, of course, a huge training ground for AI because it's quote unquote publicly available. Um, I say quote unquote because most people probably wanted their pages to be publicly available but to people and maybe search engines, but not so much AI companies uh that then make money with a question mark whether they actually do, but yeah, make money uh based on that kind of data. And Spanish is second, uh I think around 6% of German is a close runner-up. So if you're a German or a Spanish speaker, yay, you are among the top three. But remember, it's like English is almost 50% and Spanish and German in the in the single digits, right? It's a big gap. But the US-centricness also shows in how products are being rolled out. Like when Google Meet um offered live machine interpretation, the first languages that were supported were English and Spanish. Well, guess why? Because those are very important languages in the US. Um, and if you think about languages that are maybe less popular, it becomes really interesting because in those under-resourced languages, almost uh the only text available is religious and not just any religion, but the Bible. It's very easy to train uh a language model on the Bible because you have very clearly separated parts of text that neatly align across different languages, right? Because all the verses are labeled. What other book or piece of text has that built in? And then you have Jehovah's Witnesses magazines. So yeah, hundreds of African and uh Asian languages kind of rely on that kind of text, which is of course not um native to those cultures at all, but still, if you use AI in those languages, it's kind of that's you know, that's the training material that the AI works with. So yeah, to answer your question from that point of view, what parts of human culture I think? Protestant, North American, most likely.
SpeakerAnd male, of course, not to to um to drop that entirely. There is a huge gender bias, of course, in in the language. Um, and the language is how we how we interact and how we perceive AI, and of course, that also reinforces the gender bias again. Um the the challenge is also that if you as a user are very aware of of those biases and you prompt your AI to drop all of those biases and work differently, it's very, very difficult to consistently um steer it away from its own default settings. It will continuously drift back to what it assumes to be the norm. Um so uh it's incredibly frustrating and hard work to not have it creep back into your work all the time.
Speaker 2It kind of shows in small things already. I can't even tell you how many times I've told my AI systems to only use British spelling and British everything. Punctuation, lexicon. It's really difficult. Like if I don't remember to build that into the first prompt, it always defaults back to American English. Even if I if I have it as a system prompt at like instruction level, it will still find a way to bring back all sorts of things that I don't want to have. Like I keep telling it, I don't really want the Oxford comma. You keep telling it, I don't want to use the Z, I want to use the S's. And still, um, I think it's too strong a signal from the training.
Speaker 1Yeah, that's really interesting. And that's what I'd like to talk to you more about is being very clear about who you are as a brand, you know, internally having these guidelines. But so many companies and organizations don't have that clarity on how they speak internally, so that they have some clarity on what's acceptable to be communicated externally. Do you have some thoughts on that?
Speaker 2Yeah, I think uh you've pretty much um hit the nail on the head. It's very noticeable that AI amplifies this kind of problem when a brand voice only lives in people's heads. And I think some organizations have a leadership model where a leader gets away with saying, I'll I'll know it when I see it, and they become the person who decides thumbs up, thumbs down. You can't hand that to a model. People will use AI nonetheless. I think Holger has a couple of things to say about shadow AI from the culture perspective. But if you if you haven't got that kind of written-down stuff that you can hand to the model, the most likely thing that will happen is that people will think, what happened to them? You show up in public. I mean, I have to be honest, like I did a little experiment uh at some point last year uh with a with a new new AI writing tool asking it to kind of create LinkedIn posts from my existing content. And I was like, this looks like I'm a completely changed person. Like, am I on drugs? What will people think? You know, have I been kind of kidnapped by aliens and somebody's impersonating me? I don't know. But this is what happens. Like if you haven't got a style guide that uh you take seriously, if you only create it as kind of tidy-up work and not as infrastructure that serves both people because it makes their lives much easier, but also helps you work with AI better because then you actually have a little bit of an efficiency gain, at least in the editing stage, if not in the writing.
SpeakerYeah, and uh Sabine mentioned the shadow AI, of course, which is related to guidelines, right? The best guidelines in the world that you have around how you want to communicate internally, but also how you want to use the tools and how you want to manage data privacy across your organization, etc. They're not that useful if people still bring their own AI into the job. So uh it's it's very difficult to um prevent that from happening unless you educate everyone on the risks involved. Um we have seen this in organizations where even for us as a supplier, um, we have to sign up to very strict guidelines on what to use and what not to use. And basically, we use their systems when we work with them. Um but then every now and then there's a message that pops up that's like, oh, I um did that with a completely different non-official AI model, and this is the outcome. It's like, oh, so what was your prompt? What data did you share with that non-official AI supplier, etc.? Um, and this is a tricky one because we see lots of advice about having diversity of opinion through the use of competing LLMs to achieve certain outcomes, and technically speaking, that's a good process to do, but that all has to be governed by the usage guidelines of those LLMs. It can't mean that everyone brings their own AI along in a personal or not even or even free subscription, and then company data leaks into all those models in a completely uncontrolled way. So that's one huge problem. Um, the other problem to do with language and what becomes acceptable in an organization for me is like one of the hallmarks of AI for me is this incredible creativity and avoiding and evading accountability for mistakes that AI just has become so poetic in. Um, I made a mistake, mistakes were made, but I'm not going to call it that. I'm going to wrap it in a way that doesn't even sound problematic. And it clearly doesn't sound like I owe the mistake. It's basically, yeah, well, I've sort of, I mean, I don't even want to start citing those examples. It can be funny if AI does that, but when that sort of evasion of um accountability becomes acceptable through the tools that we use, and it kind of starts to become acceptable within like humans and colleagues acting that way, which I think it will because we we adapt language. I think there's there's a clear there's there's there's certainly clear research that DAI language is affecting our language again. So it's basically we become copies of what we see and Act with every day. And if that then extends to behavior like that, that's certainly a culture. And I think the last one, probably related, is what AI, particularly in terms of copy, is really good at is creating content that sounds very good and convincing but lacks any logical structure or architecture. So it sounds good when you read it the first time. It still doesn't make any sense. And I think it can also lead to the acceptance of glossiness without substance across companies. So something that marketing teams have to fight the perception of for many decades of basically only spending their time and money on things that are glossy but lack substance is amplified through the use of AI because now everyone has that problem and everyone produces glossy stuff that doesn't really have anything deeper in it and doesn't often make well often doesn't make sense either. And if that then gets out on websites without any human editing, and people say, ah, it looks great, but your customers may not make any sense of it and then don't convert, well, clearly that's an issue.
Speaker 2Yeah. One thing that we often talk about is what's what isn't clear internally can never be clear externally. And that can also be reversed. You know, it's it used to be very obvious that uh companies that were communicating in a chaotic or not particularly aligned way on their website or elsewhere, that there must have been a problem on the inside. But now if you have AI generate all your stuff, it's very easy to hide the fault lines on the inside and act as if they don't exist, even though they are a tickling time bomb, and you need to really work on that kind of stuff and produce the content from the inside out from scratch, as it were, um, in order to bring in that alignment. So it's like it's fake alignment, um, which is also kind of related to a cultural thing going on where we talk a lot about efficiency and doing things fast and for minimum cost, maximum return on investment. But we're not really talking about the effectiveness of what we're doing. What change are we actually producing? Um, both for our organization, but also for the for the bigger uh uh society that we're in, right? So uh effectiveness and efficiency are not the same thing, even though they read very similarly. And I wish we could find a way to refocus on really making a difference over just saving time and and effort.
Speaker 1Yeah, I mean, that rings so true. It's almost as well that people have forgotten that when they talk about you know social value principles perhaps on their website in order to attract a tender or a big project, that there actually needs to be ingrained within all of their communication because if there is a mismatch or someone just feels there's something off, that's something that you can't like spot yourself, perhaps, but other people can. Do you have any thoughts on that?
SpeakerThat's an excellent question. Um and I think it's it's it's social value principles that you refer to, um, but it's anything to do with with cultural values and purpose that that organizations like to use. Um because um our whole objective, or one of the main objectives of our work is making those principles and value statements observable uh in an organization, which means I don't actually care much about the poster on the wall that prints something nice that I want to read. And in the same way, like I don't really care much about the social value statement in your tender application if it's not real, if if I can't see it. If if if the actions of your organization I perceive in other areas contradict that value, then clearly it's at best not observable or it's deliberately misleading in the places where you think it adds strategic value. So we like to just call it observability because it's less judgy. But um what what it means is those values they they're not just employer branding or or something that you do to purpose dress your organization. They need to determine the outcome and the way you make decisions in your company. They need to they need to be part and parcel of every interaction that people within the organization have with one another. Um we we like to call it the operating system on which your your business runs and makes decisions. That's culture. In the end of the day, um everything you should do and you don't do and the reasons why need to be traced back to the core values and principles that you define for your business. Um and honestly, that's where the issue usually starts because despite everything around culture in the last couple of years, uh, and also all the DEI work that's going on, a lot of organizations think that culture exists and we just have to extract it into labels and print it on posters. But the work is actually to first analyze what is the undocumented culture in a business, make that explicit, and then see what works, what you need to hone and evolve, and the things you need to drop because they're actually not helpful. They exist within your culture, but they're not positive contributors, they're just behaviors that have organically developed out of certain dynamics or phases the company went through. But it's really important to get a clear understanding of all those components that exist in any organization today in order to then intentionally shape those. And that's how you create the alignment that you asked about uh in the beginning. This is where you can connect verbiage, um, value statements and purpose statements to actual observable reality, not just within the company, but also outside of the company. And um, that's where a lot of the employer branding initiatives or um uh efforts around corporate social responsibility, etc., fall short because they're just dressing up exercises. They're not actually deeply ingrained structural strategic initiatives to really penetrate the entire organization at a deep communicational level.
Speaker 2And just to bring it back to AI and give an example, um, we are a queer-owned and queer-run business from scratch. So I think it's worth saying because that shapes a lot of how we perceive what AI tells us. When I use AI and I don't tell it what my values are up front, it will proceed to tell me all sorts of heteronormative stuff that is ultra-capitalistic and not really um helping to bring about uh anything to do with diversity, equity, inclusion, or belonging for anybody. So it's it's it out of the box. I've found it ableistic, I've found it ageist, I found it, I don't want to say racist, because I think probably Silicon Valley has made uh more progress in the last five, six years in that area. And also I'm white, so I'm less likely to be able to spot it. Um, but definitely when it comes to sexism and heteronormativity, I've had my fair share thrown at me. So I had to come up with like a way of prompting the machine that tells it these are my values, and all your answers need to fit in with that value catalog, otherwise I will be able to work with it. But I think a lot of companies don't have that kind of values catalog that they could even feed into the system. That's part of what we mean with infrastructure and operating system. If you don't have that, your AI will gradually shift your actual culture in the direction of the origin of the models it was, you know, it's it's operating on. And so I don't know if you become more like anthropic or more like open AI, but you certainly become less like an organization that's um, I don't know, based in Sheffield and a traditional business, or based somewhere in Bavaria or in the Czech Republic or somewhere else. You know, it's like we lose a lot of the diversity that makes businesses unique to where they are and who they serve. And everything becomes very beige and samey and also less friendly to actual human beings with their needs and their perspectives. And that's my my biggest worry when we don't really make uh values important enough to document them properly and unlearn a lot of the culture that um creeps in that we never actually intended.
Speaker 1That resonates really with me as well. So if people are now thinking, how could they work with you? You know, how could they go on this journey of improving their communications but also their company culture?
Speaker 2I think the the easiest way to get started is to go to our website from scratch.net and book a diagnostic chat. Uh there's a big button right on the homepage. It's the easiest way to get time with us. It's free. All you need to bring is pen and paper or a tablet, and then we walk you through our diagnostic model where you map your organization along nine different maturity areas to find out like where are things going really well, where do you need treatment, and then also find out can you do it yourself? Do you maybe need somebody to coach you through it? Because we also do that. Um, and then also maybe you actually do need or want a project to kind of do some groundwork for you and then an implementation phase that gets your whole team on board and teaches everybody how to keep going with it. Uh, those are the main engagement models we have. We also do speaking and that kind of stuff. But I think the easiest way is just to go to the website and book a diagnostic. There's no strings attached, although, of course, um, if it turns out that we're a match, we'd be very happy if something came from it.
Speaker 1Brilliant. Okay, and now it's time for our mystery question. So Kirsty, um, who joined me on the um podcast last time, has left a question for both of you to answer. So the question is what is something that was said to you in the last week that helped you?
SpeakerThere is actually, yes. There is a quote, and it's not something that a real person said to me, hardly because we were pretty busy last week and hardly met any real people, other than those on video calls and webinars. Something that I I saw on one of my favorite shows that was said then again by someone in uh a conf at a conference that I attended, which is um the Ted Lasso statement, budget doesn't really exist, it's just a made-up number. That um I think, and of course, that sounds very salesy, and I will promise I'm not going to throw that back at a procurement person I talk to. But I think it's important to realize um that we we often refer back to big constraints that we feel exist and give us reasons not to do things that are completely self-made. Like a budget is a planning tool to help keep things in track and control. It shouldn't stifle critical initiatives within an organization that could make or break the organization long term just because you ran out of it. It it's it's something that you can control rather than something that controls you. Um and I think like why it resonated with me was not because of budget, but simply because of the mindset of questioning the default that we often uh don't question. And it can relate to anything we deem to be fixed and static and not changeable.
Speaker 2Reminds me of the quote by uh Gay Hendricks: You're where time comes from. It's a similar uh point of view, right? Like time is also, you're not given with a time uh budget, you're not presented with that, you you created yourself. Good. I've had time to think about what happened to me last week while Oliver was talking. So thanks for going first. Um, I think what stands out to me most is that I received a lot of questions last week from uh people who I did not expect those questions to come from about things where I thought, oh, I'm sure I was sure they knew this. For example, um, somebody who I deem an expert in SEO asked me about something that I work with, uh with which is conversions. So on a website, you want to want people to do the thing that you want them to do, and if they do that, you call that a conversion. You can measure it. And so this person asked me, So I have a project where it's also about conversions. Like, how do you actually measure them? And I would never have expected that somebody I've known for like, I don't know, more than five years who is really into things would ask that question. And I'm not saying that to expose that person, but just to remind myself, you know, this is like the curse of knowledge. I seem to think that everybody knows what I know, and I completely forget what other people don't know. And that also becomes a problem for how I relate to the world. Because would I go and explain to people how they measure conversions? No, because I feel like everybody already surely they know this. But clearly, even people who I think are very smart don't know that. So maybe I should talk about it more. So that was really helpful, and it was not the only question I received that went in that direction where I was like, oh, I thought everybody knows this. Clearly not. Good. Thanks for the question.
Speaker 1Very good. It also just shows that things move at such a rapid pace at times that you sometimes can't keep up, you know, with all the technology. That's why it's important to have other people on your team, like yourselves.
SpeakerSo and it's also important to admit when you don't know something and just ask the people who might know it. I I remember that last week we had a couple of confused facial expressions that we exchanged when lots of new acronyms suddenly popped up during presentations, and we were like, what does that mean? Ah, that's okay, thanks. Oh, yes.
Speaker 1Oh, brilliant. Uh so what question do you have for the next guest?
SpeakerWe have one. Yes. It wasn't easy, um, because it's always hard to come up with a good question for someone you don't know. Um, but I think we have one.
Speaker 2Uh, in keeping with today's episode, we'd like to know what's an unexpected way you use AI personally and professionally. Doesn't have to be the same thing, could be one thing personally, one professionally. Um, but it would be good to have both sides of the story if it if it exists.
SpeakerAnswers on a postcard.
Speaker 1Lovely. I love getting a postcard. So send one to me too, please. Lovely. Thank you so much for today. I learned a lot. I really enjoyed the discussion, Zabine and Holger. Um we'll definitely keep in touch. And for all the listeners, please connect with Zabine and Holger. Um, you know they are a wealth of information and such wonderful and creative people to have in your team as well. Thank you.
Speaker 2Yeah, thanks, and uh thanks for the great questions. Um, a good question is worth so much. So thanks for coming up with great questions.
SpeakerThank you so much.