Uh First of all, greetings from Switzerland from Zurich. It's a beautiful uh late morning here, a bit rainy, which is not uncommon in summer in Switzerland. There is a joke for all the experts who come to live in Switzerland that how beautiful Swiss summer is.I love it this year it was on Tuesday. So um I'm a founder and the CEO of Nanos. Uh Nanos A I is a machine learning start up for online marketing. And uh my background is um I'm coming from corporate research. I used to work at the Walt Disney Company and was leading a team innovation team for um creating technologies, machine learning technologies A I augmented virtual reality for the one of the largest entertainment companies in the world Walt Disney Company. So after 70 years being in a corporate career, I came up to the idea of applying machine learning uh algorithms to online marketing and came up with the idea of nanos, which is a self service platform for anybody who doesn't have much knowledge about marketing to very easy to create and place online advertisements and have online visibility.

So uh I would love to share my screen because I have such a beautiful deck. I prepared for you guys and it would be such a waste if nobody can see it. Ok. Um Nevertheless, I will continue. So uh my background is, is in math uh finance and law and I possess about 20 patents in multimedia visualization, machine learning and also online marketing technology. I have fabrications in uh um I triple E infos. Uh IA I and IC it's probably these names don't tell you much if you're not a computer scientist, but they are, I can reassure you, they are very prominent venues for, for uh marketing or for machine learning specialists. So what I would like to do is to give you a hands on overview, non tech, 30 minutes monologue with Q and A included where we could chat about product about data, hiring patents, marketing stocks, fundraising and exiting a company pretty much key secrets, uh take home messages of how to build and run a tech start up.

So during this session, we will just answer those questions. How can I start a machine learning company if I have no technical background myself? What is machine learning exactly? And why there is so much hype about it by large corporations, start ups investors and customers, how to figure out if my future product can be developed with just software? Or is there a strong need in machine learning technology specifically or what to do if I, I have no previous data to start with, to create a product, how and where to hire the best machine learning talent that I can afford for my new awesome tech idea. And also not very uh let's say very uh just recent development, how pandemic has changed the way artificial intelligence and machine and start ups used to work harder and raise money. So um I would say um I would keep on going with the Yes, I can see, I can clearly see that you guys from the audience can really help me to share my screen. Um I'm happy to send you uh the deck afterwards. Uh If you connect with me on linkedin uh Sasha Schreiber. So um let's imagine you have uh an awesome idea for your new company and you will, you clearly see it will be a machine learning company because you want to use machine learning technology.

Maybe you have a connection with the university, you plan to hire students or simply you realize that you cannot really create your product without using machine learning technology. So this would be the first very first fundamental question you have to ask yourself. Can I build my product with just software engineering production with software developers or do I really need machine learning algorithms? Because nowadays, everyone claims to be in A I space. But what really makes you a truly tech company? So one of the key um differentiators would be if you plan to patent your product or any features related to your product, which is not UX and which is not code because you cannot patent code if you talk to a lawyer and they'll tell you there's nothing to patent there. Most likely it would be very difficult for you to call yourself a machine learning company. It might sound very trivial but it is exactly what it is because when you uh develop a technology, machine learning technology, it is very common for larger companies who have large budgets to patent this technology to make sure that they secure the P space for start up smaller start is rather uncommon.

And it is a decision that every start start up founder has to make whether if they should delegate certain budgets towards uh patent creation or not. And I will be talking about patents later. But if you, when you're talking to the L and they tell you as well, there's just the code and you can't really find anything novel in your idea or concept, then most likely it would be also very difficult to uh to um project this idea on to your customers or on to potential investors in other uh potential differentiator would be uh it's important to work with universities when you build a machine and a company.

Because you can, you as a founder, it's very difficult to keep yourself up, update up to date with the all the recent uh industry um technological development because you're so focused on the industry and you focused on tracking what competitors are doing, it focused on developing your own product and how it looks and feels for your customers, but really making sure that you're up to date with all the research that is happening in your field.

It is where students come really handy. Um So these are two probably key differentiators where you can already say, OK, uh I am, I can call myself a truly tech company. Um More about uh these topics you can find on my youtube channel, Sasha Shara's youtube channel or also on my linkedin uh social media profile. Uh Next topic I would like to touch upon is product and data. Uh building a product on top of a prototype. This is what lots of founders decide to do early on and this is a decision that might fire back uh in very different places later down the road, which can be really painful. So how to avoid um mistakes which I have done. And also many of my colleagues uh start up founders have done as well. So when you're building a machine learning technology, you start with the research code and the prototype, which then you most likely will be keen on development further because it's, it's such a pity to throw it away and being forced to build something from scratch once again.

But I would definitely recommend you to consider this because building a product on top of product that will require a lot of efforts in sustaining it. It's almost like putting a band aid constantly uh onto your, onto your new product. And it's, it can be really, really costly in time and efforts and um hiring additional software developers in order just to keep this whole, very, you know, very lively creature uh alive and introducing new features would be really also very, very difficult. So if you're planning to build a product on top of our existing research prototype, we have to have two very important things in place. One is a very experienced product owner, who knows how to handle a research code and uh production uh and launching it uh in a timely manner. And also you need to have at least one very good senior developer or work with a very good outsourcing company who has senior developers who have dealt with similar uh challenges in the past. Um Another very interesting uh topic which uh a question which I'm being asked a lot is um I want to build a product powered by A I but I have no data what to do. This is something we also have en encountered with NANOS.

Uh when we started our product, uh there were just not so many ad campaigns we could get our hands on. And it was really difficult to see if I actually our machine and algorithms work if our models work or if they need uh additional tweaking. And then when a real campaign started to arrive, we so OK, we really need some serious uh apply some serious techniques here, how to tweak to make sure that our models are a bit more generic. And uh what practically what we did, it's a call start problem because if you don't have much data, then it's really uh you, you, your model is only as good as, as your data is. And if your data is dirty or you have a complete absence of data, then really uh you cannot really produce good results for your customers. And this is something what we also have to do is build called uh uh start ad. So we practically had to stimulate a creation process and create tons of tons of different ads for different um industries, different products in order to feed our machine learning algorithms who are really hungry. Um Next one topic, very important. One is on hiring because your concept, your idea, your product for your product might be very, very good, but it's really up to the team that builds it, that is your product is so dependent on.

So how to find the right tech talent and how to also to make sure that they stick around and how to compensate them fairly. And also what to ask when interviewing. And uh also sometimes I'm receiving these questions from a fellow founders who are saying, look, I I want to build a machine learning product. But I'm really not a techie myself. And is that OK, how about how I'm going to hire technical talent and make sure that they produce good results. This could be really truly a challenge if you are not a technical founder itself. And this is why it's crucially important to have another key founder like a CTO or head of product who are uh technical people who could, who could help you to uh explain your concept, your ideas to the develop development and to the machine learning team to make sure that the product is developed at its best.

So having a key uh fellow co-founder who takes care of, of this, all these technical questions including technical roles. Hiring uh is key at the very start. Now what to ask when interviewing technical talent. Nowadays, we are competing as a small start up based in Switzerland.

We are competing with all large players in the uh in the world pretty much because now with the remote home office um being introduced slightly over a year ago, of course, we gaining access to additional talents across around the world and we don't need to be in the same four roles anymore.

But at the same time, of course, we start competing with all other companies who are hiring tech talent at the same time. And what at once we had lots of uh applications, I believe almost 500 or 600 when we would advertise machine learning uh engineer and for software engineers, even uh in several thousands, when we would post our uh job postings uh for hiring people into Switzerland. But when we started hiring remote, when everybody started hiring remote, uh the number of applications actually went almost down to zero. So that means as an employer, you have to stand out in the, in the this sea of other job posts and also you have to be able to provide um um a very uh fast and um diligent interview process to make sure that candidates good candidates don't drop out in between. And also after you have hired to make sure that they stick around, these are all the challenges that you will be facing as a um a new start up founder uh in the machine learning space. And again, compete with a lot of big players like Google, Facebook, IBM, just Amazon just to name a few.

And it's really uh up to uh up to you um to make sure that you introduce your company at its best, whether it's your website or whether it's your own social media profiles. When you talk to the candidates, they usually check uh uh thoroughly with whom they are, uh or they should check authority with whom they are about to have an interview. So this is all something to be really taken care of as part of your employer branding per se to make sure that you stand out and um appealing for candidates for remote candidates around the world check talent specifically. So in terms of the compensation, we have built a metrics for uh remote check talent that will uh actually very transparent about. So everybody at the house knows what everybody else is earning. And this um compensation metrics is consisted of several uh several blocks. One is salary compensation itself. And uh it's being built based on um salary based uh based on your location uh based on your a little bit on your experience and education in the case, when it comes to machine learning um field. And uh we, what we would like to do, we were uh pretty much uh following this one rule, we wanted to have the same buying capacity, let's say for software engineer in Switzerland versus a software engineer in uh Ecuador as an example.

And this worked for us really well once we build this matrix. So let's say if nominal salary for a software developer is 100,000 in Switzerland for start up, then let's say for Latin America is pretty much half of this price. And if we go to Asia, it could be also a bit more lower. And then additionally, we would definitely uh distribute stock options. Stock options is a very important part for making sure that tech t sticks around that they're interested, that they feel co ownership of the company that they're interested in staying with you as long as possible until you get there. And this is something that also uh needs to be taken care of from the start, uh putting together um employee stock option plan for your employees at least at 10% of the equity. And I'll be talking about that later. Um So it's probably quite uh interesting to also learn how uh how to uh what to ask um uh during the interviews because um during uh the interviews, uh technical interviews, I will not be covering those because that could be uh covered by your key tech uh CTO or head of product.

But when it comes to really understanding whether if this person is right, cultural fit, uh if the person has enough experience and has the right attitude to join a smaller company versus for instance, going uh for a larger company and work in a large company and, you know, have all the kind of social benefits and everything else and um certain stability which start ups cannot provide.

So, uh I ask a few questions which helps me to understand, you know, how serious candidates are about the about the work. First of all, I'll try to get an understanding whether if they did uh background check on the company and also on myself because that means the diligence, they did their homework, they came to the interview prepared happens a lot. When during the short phone call, I have usually a request um after the candidates send their first um information about themselves. It is really uh not so many actually do this diligence and that means they were not very serious about joining the company for them. It's just yet another interview and they're shopping around for better conditions to join and then maybe uh hope on another job three months later. So this is something you also don't want you and serious candidates who took time preparing for, for the, for the interview. Another question which I asked a lot is, which is really helpful. I asked them where they see themselves in the next 2 to 5 years, whether it is uh research environment academia or is it industry or maybe um is it a corporate research? Because when it comes to the tech um talent, these are the three major uh uh um path they could take. And if, for instance, a candidate tells me yes, he would, he or she would really like to go back to academia.

Um Then, uh for instance, Ana said it's possible we have phd students working in our team uh doing their P CTO while contributing to the product. But it's rather an exception, not many start up can provide that. And uh usually if a candidate tells you, they would like to go later on for PGO or for ma that means they will be away from your company for a year or more. And uh you know, life cycle of a start up can be very short and you need to maybe uh prioritize uh towards candidates who have different ambitions for instance, to stay in the industry. Same goes for uh candidates who would prefer later on to switch to corporate research, which means they would love to go to, let's say Google and other large companies and receive double or triple or quadruple of their salary, what they would get with you. And uh that's maybe their major goal. They just want to get some experience again. Uh It's totally fine uh as a candidate uh to plan this way. But for you, if you have really um small resources in terms of time and effort and guidance of junior machine learning engineers or even software engineers, you would be thinking twice.

But if you should hire this intern and uh devote some enough time to make sure that they grow and um and deliver something useful for the for your product and then depart. Six months later, we had those case cases as well. Again, more on hiring and how to compensate a remote tech talent and what to ask when interviewing tech talent uh could be found on my linkedin profile, my articles and also on my youtube channel as well as on my website Sasha schreiber.com. Um I'll be continuing with the next next topic. Um Let's dive in a little bit about into the stock options for your employees. Uh Imagine you have created a company you uh, have set up, um, equity uh, of 100% let's say it was co-founders. And, uh, I've seen lots of cases when co-founders early on decide to go 5050 let's say it's two co-founders friends. And then they said to go, let's do 5050 start a new company. And then one of them ends up doing all the work and another one takes on the, a, a side job or even a day job at another company. This happens quite more often than you think. And this is something I would would also recommend you to avoid because this is can be really drastic and really kill your company because you give away so much equity early on and without uh perspectives to get it back.

This is why uh um a shareholders agreement has to be in place where you could discuss all potential case scenarios uh for people living in your company to make sure that you keep your cap table. Clean. Cap table is a spreadsheet where you keep track of all the shares, everything is owning and a price they have received it and has to be updated also frequently. So for the end of employee stock option compensation, I would recommend um set apart at least 10%. This is shares equity which you in a way should have to contribute yourself. So you pay for it first and you give it to your employees uh of early employees key employees for free at the beginning. And then you have long enough cliff period just to make sure that they stay with the company that the right person, uh, that they were the right person to hire. And you give them, uh, um, enough equity, you can always give them just a little bit of equity and then give them a little bit more, let's say 20,000 now in stock options and then another 20,000 a year later. And it's important also to have a slower vesting period. Uh, not more than 2% a month just to. So each month they get a little bit more equity list, which they can not really sell immediately.

So this all needs to be uh, um regulated with the shareholders agreement and employee stock options agreements. These are two very important documents which you would need to have. Uh, hopefully we will be able to hire lawyers to do. So. If not, there are plenty of course templates, um, plenty of templates, uh online or uh, um, in special indicated uh sources. So, um, in terms of the um, stock options, if there are any questions, I'm happy to answer those again. There is very, uh, uh long list of articles on these topics on my social media profile on linkedin. Please have a look and ask me questions if you have any. Um, next topic I would like to touch upon is patents. This is something that really uh is important. It comes important when you're building a technical product and you want to secure the IP space to make sure that computers don't overstep your, uh, your those. And also, um, it could be a potential attraction for investors, although not all the investors uh really appreciate patents.

Those are usually quite costly and sometimes hard to dispute. So which means even if you hire the, the, uh, if you file the patent and somebody uh, is uh, um, building exactly what you have built already or planning to build, it's, it's really costly process to, to fight against it. So it's really, um, uh, we have to decide for the company have to see uh whether if you have um, distinct budgets, we're talking about up to $20,000 to file, to cover, to get the protection just in the US. And if you want to have European countries, each and every country costs additionally, but what you could do is you can file provisional patterns and for that, you, you just need to put together a very simple draft which consists of six distinct modules. One is, what is the current state of the art of the industry? What exactly your solution is? How is it different from competitors, how exactly how you built or how you're going to build it and a couple of others. And then with this write up, you pretty much go to, um, lawyers and they put together a final draft with the claims and file it for you. And provisional patents are usually very much, much cheaper.

Uh however, they only uh valid for 12 months of protections and, but they also help you to start at least patent apply process and you can always modify those, which is also can come in very handy because that's happens a lot. Also when you build a machine learning product that you deviate from your original idea or how you do it, uh um in order to, to build something that is viable, again, patents are very expensive. It is you have to think twice whether if it's worth it, fighting uh patent infringement is expensive and um filing a provisional patent, it is something that I would definitely recommend. Um Although if you don't file after 12 months, uh the final patent, then of course, your protection also expires. So it would be a little bit of the budget wasted as well if you in the end decided not to file your, your cool idea as a patent. Um More on the patents again uh on my website such as shaper.com or uh on my social media channels, youtube and linkedin. Next one is a very cool topic is how do you mark, do marketing for your machine learning start up? And it really depends on two. Actually, it depends on whether if you're building uh A B to C product or B to B product and it's really cool and it's a short term trend to build B to B machine learning products.

So uh product that caters to larger medium size companies, uh B to B business to business and um investors love this model. So there are a huge cohort of investors and B CS and family funds who only invest in B two B and don't invest in B two C for various reasons. Uh because probably they feel it's, there's much more potential to get a higher return on investment. That's probably the, the only um uh assumption they have. But uh I have to admit, in my opinion, it's really short term, short term trend because what we see also at NANOS, um if you really look deeply direct to consumer tech is really picking up all various, all kinds of various um self-service platforms are now picking up. And there's also now number of investors funds are growing who invest particularly only in B two C direct to consumer technology. So, so that would, that is a smart way of dealing uh planning your product road map. So first you are catering to B to B companies and then you branch out and use part of your developments and technologies already produced uh towards direct to consumer uh product.

And when you are um building a product for B two B, the way you start marketing is, is uh definitely um you cannot just go out there, place ads on Google and Facebook and Instagram and expect B to B leads coming back at you and buying your probably rather uh rather on expensive side, probably with some subscription service.

So which means you have to build up a sales team and also leverage um professional networks like linkedin and others. In order to reach out directly to your potential buyers and this makers in the company, there's also longer sale cycles, of course, uh when it comes to B to B um products, marketing of B to B products and also very distinct content strategies. So we have to build up a lot of materials, presentation decks. We at NANOS have about 20 different versions of the product deck for our B to B product white label solution of our NANOS of NANOS. And um it is something that also takes a lot of time and effort from your team or yourself. If you're um if you're still a very, very small team uh in terms of the direct consumer tech, it comes actually a bit in a way a bit easier uh because uh you can really rely on paid marketing P PC uh campaigns when you advertise your direct consumer product. Because your direct, your customers and customers, they are on Google and Facebook, they're searching for services similar to yours.

The I will find you if you bid for the right keywords, same for um same for Instagram, your uh potential customers uh and customers most likely on Facebook or Instagram or on tiktok. So it's rather easy to, to get to those uh to get to their attention, to bring them um a knowledge about your product. So, uh again, nevertheless, it is always a hybrid of really good content strategy, producing uh materials, articles, videos, short videos, product walkthroughs that uh needs to be built for both uh for B two B and B to C technical products. Next topic would be fundraising and exit.

How easy would it be to raise money for my new tech idea? Uh I have to say upfront, it is now much easier than it used to be. So that's the good news. Uh The bad news. There is so many obviously tech different technical uh concept and ideas floating around and some investors are getting uh also technical savvy and I will get uh technical advisors on board which is highly recommended for investment fund to have if they invest into technical start ups is to actually have technical advisors who could screen all the start ups and to understand better.

Uh how far fetched the idea is the planning, the start up. Um The given start up is planning to build. So um I was told raising money in the US is easier than raising money in Europe. I would have to disagree. I feel there's a lot of money currently that is being allocated for tech start ups in Europe. I'm not talking just about Switzerland only, which is fantastic uh start up ecosystem for deep technology start ups uh in specifically, but also in UK and other countries. Also Germany.

Nowadays, there is really, there are long list of family funds we c investors, uh angel investors who potentially uh are willing to search for this for the next big new thing. Uh There are four types of investors which you could reach out to for your technical start up ideas. One is, of course, your start could be self funded because if you just created a prototype and tried to reach out organically to as many users as possible, if it's a direct to consumer product, I've seen these cases in the past uh uh companies reaching out 10,000 users uh with the just with the research prototype.

So it's possible to do without any marketing budget and even the company being created um If the service is free. Uh So that's one way of it's completely bootstrapping your idea before you actually um agree among themselves that you're up to something really big. Second uh idea.

Se second type of investor is uh um family fund. Uh Those are also fairly easy to reach. Usually they have a um simple landing page, easily searchable online. And also they usually um keep an eye on various uh start up auxiliaries and so on. So it's, it's really nowadays easy to get um just simply reaching out to them on linkedin. Uh with direct message also works. I did this in the past and it it it really works nowadays. Uh now during the pandemic, everybody is online and very active on social media. Uh Next type of investor is of course angel investors, various angel investors, former start up founders who um build companies, exited companies and now looking to reinvest their assets into uh uh ideas of other people. So this is something also which is really interesting uh for a fellow founder to explore just to connect with other fellow founders who are a bit more successful and try to get the recommendation to other family funds or to themselves if they would really be willing to invest money into your start up.

And of course, the last uh but not the least is VC funds. Uh which uh at early stage, usually they uh not very keen on investing unless they see some serious traction, which means up to 1 million recurring revenue, annual revenues, which is sometimes quite hard to reach. Uh if you're just starting out is uh really out of reach for most of the companies. So this is why it's not really an option. Um Another important topic is when is my company ready for sale? Um It is really uh interesting because I've seen so many cases, tech start ups being sold really early just because they had one patent that that particular larger company really was interested in. Um Sometimes it's uh sales happen, exits happen. Uh when um strategic investor is invested into the company by strategic investor, I mean, um a larger company who is acting in a similar field as yours, for instance, uh let's say um um a good example would be uh optimizing um something marketing technologies for ecommerce and then being bought by Amazon who would be potentially a also strategic investor or potential buyer per se.

So this uh comes with disadvantages if you planning on selling your company to a strategic investor or uh a larger company who is in your field, uh there is a lot of uh um hidden um underwater stones which need to be uh uh really looked uh carefully on because uh sometimes strategic investors depose um different conditions uh limiting conditions on to companies.

Uh One of them could be, for instance, that this particular company could only be sold to this particular strategic investor. It happened to one, one of my friends, one of my colleagues and it was really, really painful, practically stagnated the whole company because they couldn't take on investment from other companies and they could only sell to this particular free first investor. And that pretty much um they, they could not, didn't have any room to, to, to move around. So they had to be sold to the strategic investor at a very low price. And it was really, really good and exciting idea. Um sometimes uh machine learning companies uh more often than not, they've been sold because of their technology. Uh rare cases, they've been sold because of a very large customer base or a product or a tech team. So that's another type of the exit is a technical hire. So, uh with uh your technical talent, you have accumulated over, let's say the course of a year, year and a half. Um it's an early stage, uh let's say 56, tech is uh very smart um working towards a product. Uh It's a very attractive uh target for larger companies to acquire because of the tech talent nowadays is very hard to find. So I am done with my um with my pitch uh with my uh talk with a monologue per se.

I would love to hear questions if you have any. Please drop them into the chat. Um a very good question. If you work in close to academics, universities, would it be a benefit to apply for research grants? My experience with research grants, they are very time consuming and there's no guarantee for success. Um It's good to have a partner as an university where they have a distinct pipeline, at least certain success rate, you know, for sure. Uh some, let's say 60% of all the grants are being approved and we did this in the past and that works. But if you, let's say, just have some connection, let's say technical advisor from the university and you plan to apply for research grants here in Europe or um beyond uh my experience, it was very time consuming, very costly. We had to produce an enormous amount of documentation and projections and God knows what, which we could not really reuse anywhere else. And in the end we rejected. So that was very painful. Next question about patents. I love this question. Uh Enforcing patents isn't expensive.

But what is your experience with patent insurance? Um We didn't do patent insurance. What we did was provisional patents uh once or actually two times and it really paid off because it's a little bit. You buying time, you invest $500 you file provisional patent and practically you cover a little bit already the space and uh have um um have time to develop your technology and think. Really are you, do you really need to file this patent and spend uh these thousands of dollars in lawyers fee and lawyers hours and also your time because we have to explain lawyers, your technology. You have to check all the claims, all the summaries of claims and also provide additional um um um documentation if requested by the patent committee. So, um, but per se, his pet insurance, no, but I would be happy to hear your experience if you had any finding the right patent attorney in this space is also, I could not agree more. We switch three, law firms uh who had been working on our patents. We also work with uh uh with a sort of uh cheaper version of volunteers. Um uh per se, we end up uh scratching all the work starting all over with uh uh with a serious law firm again.

So it was a little bit of waste of money, but more waste of our time of our machine learning engineers. Uh I would say if you found a good patent law firm do share contacts here in the job because it's, it's really hard to find those that are reasonable with their billing hours. Those that also don't bill you all at once. That would maybe give you like um an option of uh dividing your uh payments into monthly payments. Uh That helps a lot to not increase your burden too much because this is something that also investors looking at a lot. What, how high is your burn rate? And if you all of a sudden get the $20,000 bill from your law firm and you have to fit it somewhere. So it obviously immediately reflects on your bond rate, which is not great. Uh And um sometimes uh law firms, they are also good at uh putting a lot of billing hours only to tell you afterwards that your uh solution is not really patentable because they cannot patent code, they cannot patent soft architecture per se. So you have to really be creative and turn these conversations around and uh argue that this is not a soft architecture, this is machine and architecture and there's a certain novelty and how you use machine, certain machine learning technology specifically in this field in your area or this specific feature a user need.

So this is something that also takes a lot of your time as a founder. And particularly if you don't have enough technical background, that means you have to delegate, that means it takes time of your employees or other key founders, which you could have put towards, for instance, developing a greater product. Um re read the small print. Yes, even reading the small print sometimes, uh if you don't have much experience working with law firms, uh it's, uh you know, they're always better, they always had on the game because they're lawyers, they're being paid it. But uh it is really, uh you really, there's so many things that they could get back at you at and add additional billing hours and I'm very sensitive to, to bills from lawyers. So, uh I had again, it took me a while to find the right law firm, uh which don't bill us as a surprise as yet another uh another invoice. But uh it's, it's really, it's a bit of luck. It's a bit of, you know, recommendation. Also. Referral helps a lot. Uh Our current company, we file patents with uh it was a referral from uh my past corporate research career from the Disney company. Very nice, very nice. So we are a little bit over time. So in the case, uh you have more questions.

I'm very happy to um to answer those at uh this following email address. I apologize for the um delay on my side. I was actually sitting in another session. Um, so thank you very much all of you for your um undivided attention and um hope to hear from you. Bye. Take care.