welcome to the rise of virtual chat series my name is Fabian Vesta Hyde I'm an entrepreneur and investor with a passion for show intelligence and in this episode with two guests we have a lien loja and Rasmus Jota both are the founders and the Chairman's and board of directors of the company my antics man Tech is one of the first true venture studio mainly focused only in the eye so to make sure they are building regular venture capital cases or artificial intelligence companies and we will have today a lively discussion about what a business models for AI what are trends how to make money with AI how to start and other learning everything around business development and artificial intelligence so it's it's sooo it was a one hour live show you can lock on line to see the chat and connect with the people else just enjoy and watch it see you again welcome um good afternoon good morning good evening to everyone in the world it is great to be back life again I'm
saying hello to everyone last time we had visitors from Brazil to New Zealand Singapore all over the world so as usual tell us who you are share your LinkedIn profile tell us where you're from were you based I mean we do this for two reasons we want to be entertained we want to learn something and we want to connect we want to meet new people probably everyone is still stuck at home on an office I record everything from my office office it's my it's my home from home you are sitting at home or in the office so we are here we connect we speak with each other and it's life um so today um as I said post your link and profile this is a tradition we normally do our topic today is a second so concentration Fabien um before we started the topic today I would like to point out that we have a new baby in our series so it's called the rest of my a I Academy and we're offering three courses
of the period from three weeks each course six hours so it's intensive workout including you know quizzes tests you get a certification the Annika is doing it she will say hello I already said hello in the chat ooh she's organizing together with a tech agency we're doing it for a while who really know how to you know the technical stuff a it's AI for managers so if you have a first idea what a eyes and you would like to know what tools out there what is AI what is you know neural nets and different machine learning methods go in learn it so you get a first introduction then we have one explaining Billy I this is similar to the topic we had with Beryl somatic last week in our virtual chat what we spoke about trustworthy I so there you learn the tools to understand if your AI is doing the right decision which is perfect for reports due diligence ethical discussions for your Board of Advisors you know if you want to if you
have night in AI your company you need to explain what it's doing take this course and then with a third course it's data scientists for beginners and two things I would like to add first if you interested go on our website slash Academy or just click on the menu and second we offer it for 99 years per course you know we don't do it for free because they are teachers in the background there's a lot of preparation you get a full education and it's a course for us it's a test you know what other topics you would like to and we build and we know we could have 40 courses on top of our whole period and instead of normally 300 euro per day so it's 99 years instead of 300 and seconds I will take part myself too because especially the first topic AI for managers I want to refresh my knowledge I want to see what's new out there see I have a business degree and even at work a lot of VI
I think there's something for me so if I will do it maybe you want to do it too and the Wanaka has already shared the link our guest today our last most and oddly an and they are already in the in the back in the in the green room and it will come live soon so our topic today is everything around I call it how to make money with the I so AI business models AI trends everything you know combining making money making revenue reducing cost but the AI and who's really good speaking about this are our guest today um I know when I met oddly on the first time he was doing a sabbatical so he just had sold one of his many companies it started had the liquidity took a sabbatical he troubled and then he got kind of bored during the traveling and then we met in Berlin because he wanted to do something new into AI and I was one of the people he contacted in Berlin say hey I like
Berlin I like a guy let's talk and we talked and then similar time I met that's most on a meet-up where he talked about how you use deep learning I think to mass gain a lot of faces for dating algorithms and something like hot or not just with the I which was let's say giving him a lot of bus time and shortly after this they both started near antics and we will talk about what my huntys is doing a little bit later so when I met them both were you know one from the research side doing AI the other one from the business side not doing a lot of AI and they had not really an idea where they will be five year of four years later four years so here we have a Osmos and I asked them for children pictures before and they shared them with me so um that's Moses by Education and engineer he's in he studied information engineering in Oxford and Princeton which are some really good universities and then
he did his PhD in a tad silly which is another really good University so just on the paper technical he passed the best education in this space he has like 15 publications he was quoted a couple of thousand times he's founder of HEC zero when he was still in Zurich and then he came to Berlin and he co-founded mad antics with with odd Leon so that's Moses more on the technical side on the AI he is four with 30 under 30 he was in the students shift talke and which I like pretty a lot he's founding member and board of directors of the German association for AI which is a really cool thing which was initiated by both of them to say hey we need a lobbying group for German AI groups and they think within a couple of years two years they had over 200 members already in the Association and then on the other side the other one you know handcraft kit and then with idli Aloha and Adeline logo it's more the
business side he studied in Zagon he attended similar to university he's member of intrapreneurs organisation and you know this is how he described it really well-educated business person he has started and this is how his own description 10 companies of four with four exits means he has failed before or still keeping the companies very valid if someone I like it's saying hey only 40% of them I sold which is the fear and honestly a good one the biggest exit there was over 60 million dollars I assume it was signed but he never told me so he was involved in some good exits he's an angel investor as usual and I also like what I you know did my research he sells somewhere in his files he reported to Mac Zuckerberg because and very early I think 2007 he was working for face for face book so he's in the field for a while and today and I'm sorry it's not the best picture today they are the founders and majority owners of mad antics and
math antics is in self-description adventure studio so I describe it as a company builder slash accelerating elements plus early-stage front you can read it in the news they have raised 25 million years just for their own companies they have a twenty five million fund they are launching let's say two to four companies a year they already have not a couple of a I driven companies and now even they have the money to fund this what they do they will describe this wonderful that's why I love technology right so back um they do the sourcing revelation the funding the incubation so if you're going to open or out there or you would like to start a business or your smart person this is the perfect entry if you're just a typical startup and having already funding this is not a typical VC you approach this is the one if you're intrapreneur you want to do something they provide you with data with research with regulation bring in the talent the funding and that's why it makes
a lot of sense to have a company build or adventure studio for this because in AI and we talked about this now you need more than if you duty commercial something like this that's having this set now let me bring them online you used you don't smile so my introduction was not correct all the time no it was awesome thanks very was very flattering good so um as usual for everyone out there you can use the ask the question function and I see the questions then I will go through and then pick one I will ask both of them you can address them to both of them or you can address them just to add Leon add your wave your hand for a second yes and last most your hand so you know known the street where they sitting we're talking today about AI and making money with it what is the business model of a venture studio well our business is building companies and investing in them and so in essence the business model
is around creating companies and over the course of time of course also going to exit them even that's something we believe is going to take quite some time which is part of our core beliefs why we have set up the bench a studio in the way we did with a with a firm that has over 10 years of run time we believe that you know AI is still in its very infancy even though people talk about it a lot where you still believe that the biggest impact will be seen more like in the next five to ten years to come and that's almost you only earn money when you sell the company I mean that don't doesn't your portfolio companies try to earn some money as well and if we lost I think no you have to answer no no for sure I mean like obviously I think that's also something I think we will talk a lot about today is all these companies have solid business models and are generating revenue but I think
from from an investor perspective so we have 25 million euro fund where we have some investors behind it and from them it's very much 10 15 year horizon where they don't care that much about the revenue next year the after but they care about can this company be truly impactful in the next 10 15 years and have a big impact and thereby also have generated significant revenues and so I think we're very fortunate to have these investors on board who also think very long term about this because for some other topics we go into they are in quite complex regulated markets technology is hard and so it's very different maybe from the last decade or so where a lot of them more maybe like you know putting an app in the App Store is something where you get feedback very quickly there's little regulation within a month or two you see people using it and you can iterate a pivot but for like if you bring a medical product to the market that's not something
you pull off in a month it might take a year or two and so the horizons are just more longer and so we also try to get the message behind it to kind of perspective so for my entrepreneur and VC experience I think this is a highly interested model first if you're invest in the a company it takes up to two years to build a product market fit to go live and you need patient investors because you control traction so humor antics you help them over the first two years make them ready before external investors are coming in so it's really like having them in a you know closed environment and picking them yeah I think that's a very interesting description I think our model is all about validation validation not only in terms of business models but also in terms of to technology and of course also the single most important factor of early stage which is a team and only after these three factors have been properly validated we believe a company is
ready to you know go out and graduate from from the venture studio and and you know raise external external money in the indie series a so our model we have is very much focused on providing seed and proceed but besides that and and we believe that is really important here money is not the only thing we're investing nor do we believe it's the most important at this stage of company building in deep tech so this whole environment of synergies and the ecosystem that we created which is all about you know technology platform that is shared but even more so business best practices technology best practices but also access to industry for first POC for first validation to win customers and partners is truly truly essential and so in in our eyes these are you know strategies that are acting as compounding factors in building companies and the more companies who built the more value is actually created on the platform that each new company that we create together with our founders are able to tap
in I would disagree there having built a company builder before that the synergies are at the end less than you think but still they are synergies and it makes a lot of sense and agree on the research learnings on the technical part on the funding operations I mean let the founders focus on building a product and selling it by the way founders I mean you are the forms of man antics now you have ventures below how do you screen them how do you get them like and what kind of people do you look for maybe maybe and of course Rasmus you just jump in but but maybe discuss let's discuss this quickly discuss your disagreement because I think it's really important and I also know where it comes from and and and that's why I would like to explain why I I think we're not talking about the same things because usually when you talk about company builders the synergies you talk about our shared engineering resources marketing sales so it's it's a lot about
shared people that help building companies and while we believe that this is a an interesting model for companies that are very much execution based it is absolutely not in our eyes at least four deep tech companies that are very much technology driven in the beginning but then also business model driven so we are not believers in creating some cheese by sharing resources which is mostly FTE in the individual company builder cases but our synergies are much more intangible in a way because it's much more around you know this common technology stack that you built but then have engineering teams that are dedicated in the companies and they only work for this one company working on and then also having people share their knowledge and expertise among the companies in a in a rather informal way so it's not that we try to enforce things here but they start to happen naturally and you know when you're in a in a research heavy field such as machine learning AI you know like developments are are going
on like very very fast there are new breakthroughs ever at the end you have muted yourself maybe so until we're partying back that's most continue em I understand the discussion but let's jump in for a second about the phones like what kind of people do you look for because you can't start the business yourself yeah sure and I think that's that's honestly the the most important part of our model and probably it's also gonna be the biggest factor for which will decide whether we will be successful or not to pick these entrepreneurs to build the companies with us and in general I think like there's this there's everything you look in general for and an entrepreneur and I think there are a few things which are specifically important for AI and so maybe I'll just focus on those I think like the timelines are very long the problems and so you need to be very persistent and basically be fine within an environment their feedback and you know success might take 5-10 years and you
necessarily can flip something very quickly so you're definitely looking for people who are great leaders and charismatic can you know I write knowledge but also we have this persistent to go after these like harder kind of cases I think the other one is also to have a certain depth of understanding of a even if you were the CEO so usually all foreigners we have a CEO and a cg on the sitios obviously super technical has it's very deep background and machine learning but also has you know as it's called solid teams and but even in the CEO we would be responsible or for the business side you look for people who have at least a basic understanding of AI but then I'll just kind of the intellectual capacity to drive into really complex issues because even they might have to deal with regulatory issues or they might have to I don't know sell to car companies where on the other side all the world's companies are on the other side super technical people sit next
to you and you so they also need to be able to discuss with them on the same level and also when you raise follow-up round series a series via the investors will die very deeply to what you do and so you always also need kind of that level of expertise there and we've been looking I mean we constantly talk to entrance and residents we have we've currently three and new entrepreneurs and residents we work with - you basically end up with those three we talk to over 300 people so it is quite a quite an involved process and we're constantly looking for for for new entrepreneurs and residents to join us but the bars very high like it's in the end I mean like a lot of interviews but all also what's really important for us is to get to know the person better and work with them and so for that we actually design two days we do a two-day on-site challenge where we basically work with them together on a potentially new venture
case and have them out subject with a lot of people to really see us as their culture fit is this a person I would want to work with for the next five to ten years because I think that's that's arises we need to look at all okay so being hired for a job I mean just spec in I am back in sorry good that's if you have two partners that's perfect question for the founders what is like the the H I like 25 right from University or like 45 doing the third career I think it's a mix so some some relatively fresh off university some have right now if four or five years of worked friends have started a company before or have been a the job so um it's a mix I think we we don't discriminate on age you can also talk to people who are much older I think that's also super interesting in the end it's there needs to be something that Wow's us about the person and so far yeah
I mean that could be next yours are all like in their early thirties so I have a few years of experience okay um I flip from each two of you depending with the quality of the stream today yes so I don't tell me where can you make money currently with a I like internal like to things where as a start-up doesn't make sense to go and we say that's a good market an opportunity and what do you say is enterprises who really exists where should they start in AI like internal yeah I think there's there's is one interesting distinction that we need to make and he also brought it up already I think raciss mentioned that we were doing a lot of research in the beginning of of the company creation we realize that there are many many hundreds of interesting machine learning AI cases yes only a few of them will be valid and interesting venture models so you know in depending to start a company just yes yes yes so the majority of
opportunities we have come across we realize that this is things actually the existing industry incumbents would be best at you know implementing some machine learning AI knowledge and then actually solve it applying technology rather than you know you try to build a venture around just because the the value-add you're creating is too incremental to make for a good venture case right if if if you're looking at good at good venture cases it's it's basically you need to have an idea and I mean actual path to a let's say billion-dollar company if not it's not a good venture case right and so I think this is the first most important distinction that you need to make when you look at when you look at at cases and and also this is the reason why we have not only come up with you know building ventures but we also have our own that's a service firm which is one of our portfolio companies which is actually helping you know corporates large companies but also middle stands to
actually build machine learning solutions that then can be applied inside companies saying one of your ventures is providing knowledge and you know I know you do workshops and education too so you help clients to digitize it you say I then you also build companies which sell solutions to them so from your because you are very into the market you look at hundreds of ideas a year what I like the top three opportunities currently we all know it's ideas nothing's all about the execution what way do you see like crudely again a stretchy to get I mean I can I can start I think I think one space we are super excited about is biology at the intersection of machine learning and that is there's a lot of data and Bialik biology and there's will be more and more in the next couple of years as sequencing costs are going down for DNA sequencing and so we'll just have so much more data which right now has been not really there or very manually analyzed and
so I think there are a lot of applications for this was for I don't know optimizing proteins and then looking towards farmer application but also potentially high-value chemicals so I think this is just like a whole field which is just opening up and there are already companies out there but in the next 10 years we will see a lot of new exciting companies there that's one field and another one which is more from a technological angle in the last couple of years we've seen a lot in computer vision I think I wouldn't say computer vision is solve but like for standard image classification or object detection assuming we have enough training data we can build pretty good systems analyzing these images and so there have been a lot of companies in the last couple of years solving that for specific vertical use cases and there's also a lot of tools out there which make it very easy for companies to build their own solutions here so I think that's where already has a lot of
happen I think for the next couple of years we will see similar stuff happening and we're already seeing that in this space of like language understanding natural language understanding natural language processing where we've also made a couple of years later after computer vision kind of made similar progress like the image net moment the moment when on this famous computer vision challenge suddenly the deep learning was really strong we're seeing a similar moment in in language now and that will open up a lot of interesting applications which are around processing large amounts of documents automatically I mean all the thick customer service automation I think there is still a log they were more helping there's the whole layout voice so whether it's around translation whether it's about generating synthetic voice oh that's about dubbing stuff but also in the in the context of creating content so maybe having algorithms create reports which currently humans write based on data or you know maybe in education so I think there are a lot of or marketing so I
think there will just be a lot of application which are just now opening up because the technology suddenly slowly gets to a point where where it gets interesting for commercial use and so that's that's another area it's natural language processing communication everything Britain or the you know speaking and I can agree by the way for farmers out there this is the most mature field there are a lot of companies out there trying to solve this but no one really search then we have image recognition which is more a technology in a concept but makes a lot of sense because we humans do a lot with our eyes and replacing human senses with a machine if you have abundance of data is a very smart move and a guidette you software's pretty well so now we were looking for some interesting business models and the next one you said it's biology which is not yet there which is a cup coming opportunity which makes sense to maybe go into if you have I would recommend the
market knowledge and access to the resources because you need to get the data you need to build models and by the way will be different talk but building AI companies X expensive now are they another three opportunities from you share something from your in science well you know it depends it really depends on whether you're trying to monetize things very quickly or whether you're willing to invest and then later profit from what you've built so you know I think one one area that we are where we are excited about but we you know very little and and therefore we have not started to explore it yet but we know that machine learning will play a big role is quantum physics right so big thing to solve yes that that is going to be huge another one where I personally also am a big believer of is is the the field of neuroscience I personally see machine learning as a fantastic tool to actually understand the human brain better which is still in its very infancy
when it when you look at at our common understanding of this topic so I think there is going to be massive breakthroughs in the next couple of years might also be 10 years so it's really hard to say and then on the other hand you know like of course you can you can go into automation cases that are more obvious where you are going to create faster revenues so you know talking about I don't know ecommerce applications where you are going to automate human work that is done today at large scale with machines will also enable you to create interesting business models right so think if you talk about the ways we see AI enabling business models you could say we have two layers of looking at things one layer is the one that is talked about a lot which is automation so complex decision-making processes where human experts are involved and necessary and they're getting better and better the more experience they have that's very interesting case in general because you can definitely train
machine with much more data than any human expert in the world can be trained so it's self speaking pretty obvious that in those cases machines will be better at some point than most human experts for example like answering emails and customer support like the easy one where a packet cetera this can be for hours quick on yes yes and the other layer that is is much less often talked about but but as important as we see it is accuracy you know it's not only about automation yet it's also about how many mistakes are done in certain processes where human decision-making is involved because if a machine is trained well it's also pretty self explanatory that it will make less mistakes than a human because the machine has no conditions that humans usually are affected by that's interesting argument last week we had the topic explaining billion trustworthy eye and there we learned it's a trade-off between accuracy and explained ability and I assume regulation will lead to that you maybe have less accuracy because you
need to open you know give the frameworks provide the data you need to have less quality but therefore you can do the reasoning by the way I know that you las most are having just published in the newspapers something about this that we need to discuss this and this is coming to UM I like you answering the way that cosmos was giving us really now existing cases and you gave us the C or like visionary cases for the next five to ten years and by the way neuroscience is great I recommend checking out from Elon Musk nor in length the public demo they did and how they were analyzing brain patterns and signals and then once you have the data it's about machine learning to figure out what is a person thinking at the moment and how can I put this into a coat and sit back um you're talking a lot with enterprises what is like your I mean AI is not easy to sell currently it takes long it takes very long to
do pilots to get the data and train the systems what do you think is the largest bottleneck and interacting with enterprises to make sales to convince them to to to use AI to either build it or buy it that's a great question I think when we look at the last maybe two three years it has become easier I would say due to the fact that people have understood the concept of machine learning AI much better so their expectations in this technology how it can be applied is more realistic and they also are willing to take some risks in order to get there because you know when you when we are selling a PLC project to a large corporate that means we're not selling a waterfall project that delivers a certain product but we are actually selling a research project with a lot of uncertainty where we don't exactly know what's going to be the outcome that's why we call it POC because we first need to figure out whether it's technologically feasible you know the
business business model application makes sense and so on and so forth and only after digging in there we will be able to have a much better more educated perspective on how impactful can a solution that is built on top of this POC then actually be and and and only there we are going to be able to make better predictions how we know what what level of what what degree of automation were able to achieve how will the accuracy be this is stuff that you generally don't know when you start the journey and that's of course what makes it a bit more difficult than you know trying to sell I don't know a website where you know what's going to be the result so to everyone out that Piercy I think means proof concept and yes sorry more like if you do have a proof of design and proof of prototype a proof market product fit it's a tour of a usual you pitch to client they either they come to you because of the problem
or you contact them and tell them I I know your problems I can solve this maybe and then you get twenty fifty hundred thousand years if you're good to test this to roll it out maybe to develop a product with them as a prototype and then knowing if you can scale or not and that makes by the way a sales pipeline within my portfolio up to two years until you have a loan like monthly recurring revenue contract so do you see do you do like selling applied research I think yeah I think we have like you could bucket in like three things so like for some companies where it's very unclear way as kind as I can have the biggest potential and they might want to find out or what we still need to evaluate a lot if data is available and what you case to really focus on first we sometimes firstly start with workshops so where we just spend a couple of days with the companies really diving really deeply so it's not
it's not a very technical level really looking at data already and seeing okay is this actually possible and that's that's the first session and second stage is doing these proof-of-concept we just talked about so running small projects with them or there's still a lot of technological risk and then the third stage is building production ready systems and um we could basically enter with companies at all stages right so if if the company is still on a very high level it's just like hey we need to do something today i well can start with workshops some already have a specific problem they want to solve and then it's about a proof-of-concept and sometimes even it's already no known that it's basically possible what they want to do and then it's entirely going into a production ori system but usually then we grow get from one stage to the next one basically as we mature and i think we can see in the last color of years that companies have a much clearer expectation of what they
want so i think three years ago eighty percent of the companies where really let's do something with AI and now eighty percent of the companies already have some concrete a nice application of mine and then it's our job as possible and then built the system for that but that sounds a lot like a project agency but I know that you have put all your companies so they are having products that sell or yeah so there's really yeah we just talked about the more antiques labs you know solutions provider company besides that we have right now two other product companies vara which is a medical imaging platform which has doctors in the in the in this case of breast cancer detection and breast cancer screening to make less mistakes but also to work more efficiently so that's a medically see certified medical product in the market basically workflow solution up in helping doctors and that's it's a product we we are actively selling right now across Europe and and that's one company vara the health care
company and then there's search which is a company in the space of automated driving it's a data platform for for sensor data from cars with sensors so if you have an automated car whether it's a highway pilot or some crazy level four five so like a fully autonomous car and they have a ton of sensors that produce a lot of theta so camera data but also lighter so distance measurements and that's petabytes of data and so you can imagine when you have a few of these cars in the street you're basically drowning data and petabytes hundreds of petabytes and so our data platform AC helps to analyze all this data and make makes it searchable so you can say hey give me all the data where pedestrian that XYZ and then it will filter that and or it can also tell me what data to keep and maybe throw away because you have so much data from empty highways that you know you don't need to store it all because it's gonna cost you like
millions per year and so we basically saw this solution this product to anyone who has large amounts of sensor data from from cars okay so we have the first question by tassel to this and he asked how long will it take until an AI solution will be accepted as a commodity for I it's interesting question I will translate it a little bit like when will it not be AI anymore what's a software um I think that very much depends on the space and absolute that the topics right I mean there is a lot of things that have become commodity already and you can like like money laundry detection you know credit card fraud and something like this yeah there you have a lot of existing solutions that you can almost buy out of the box right and I think this is also what what's so interesting about about a technology like AI you know even though today a lot of what we do is is pretty it's pretty I would say close to rocket science
and as a lot of research and development we are not under the era of assumption that this won't be a commodity and ten years from now as it happens with all technologies right all technology start being very exclusive you only have a little group of people who really gets it and can actually apply it and then as things evolve more more solutions come out more and more is commoditized and then at the end it's not so much about technology anymore but about how it's actually applied to enable business cases right and this is also the way we look at at at AI as a general purpose technology that in our eyes is going to be the most important of of the 21st century but it will be no different than for example the internet before that has commoditized a lot and today you won't find a lot of companies selling internet technologies but you will find a lot of companies selling business models that are enabled by internet technologies so they say it's either X
plus AI yes so you just take anything and combine it with AI you take business more infusive pee I like you know you're an automotive company you can use self-driving cars to build in your business model around this my personal view is AI is a commodity when it's free like the internet or websites like typical word just accessible with WordPress and other similar I think WordPress is like driving up to one third of all the net internet pages out there and it's free so everyone can build its own website I think AI is finally there when everyone can apply build customize the own AI even enterprise yes drag and drop like in in a sexy SFP or do it at home like installing in an app and you have any I'm managing something for you and I agree this will come market by market but we will have this so for my view or in other questions what is currently I mean we all know that the eyes growing you you do more business I
do more business this is a growth still it's not perfect there's still a lot of obstacles out there we know that Chinese are very fast regarding funding and data and regulation the Americans have with Silicon Valley in head start you know Weiner ap' are behind what do you think what you need what could change up there which would make your life a little bit easier are you referring are you referring to to Europe like what we need to do in Europe to not loose enterprises economy as politics society I mean what are the factors currently every well bottlenecks what is currently limitation for more growth or more success and then overall in the air market so let me start with the with the overall narrative and then I'll have Rasmus go into into details you know it's not always I think a fast answer to a question would be we need to change narrative how we think and approach new technologies in in Europe because approaching it in a way of thinking of regulation first
before exploring is just not going to work you know the belief that you can minimize mistakes and failures by putting a lot of regulation upfront in my eyes is just never work and that's never worked has never worked and this is seriously this is seriously of course posing posing threats to you know the European ecosystem let's let's put it that way the European deep tech ecosystem if we're always attacking things from this angle first and at the same time you have a a culture culturally different view in the u.s. in in Silicon Valley but but also the rest of the US where you know failure is just factored in as part of the process and it's okay ai is failing a lot you said it's a research so a guy's failing way more a lot yeah and and exactly and then on the other hand you have China which has a very different view view on this and I think it's it's very interesting to see how they perceive AI as their own like almost
personal moonshot a bit like you know John F Kennedy has defined in the 60s or in the 50s already okay let's go to the moon this is going to be our first foremost national priority and we're going to do everything it takes it's the same right now with China has articulated on on AI I don't think there's any other technology that is as highly ranked in terms of importance by the Chinese government in investments hundred thirty billion investment into this I agree I read global superpowers by carefully or some blog articles about this AI is going full into it I think we'll have a whole series just about Chinese AI because they doing a lot of things right okay so you said politics is regulating instead of enabling you said I think a little bit like society mindset you know we have to have it first before we control it and then you set mindset in China political willpower and United say it's more liberal well we are just created we do something what else
like these were three what could change like politics won't change that easy mindset and Europe won't change that easy take a new track move to work and role models right we need to work on role models lighthouse lighthouse cases that really work well and can be used as an example how to do it and can be used to you know reduce the reduce to perceived risks of doing new things so your audience out there that's what we do it here you have to role models to people you can identify with and aspire to be similar and they will keep doing the things good because we need role models and honestly compared for European AI companies Mirotic's is under the top 10 according to some other sources like Forbes and hundreds blood etc so you're doing something right so we need more like this I think yeah one thing is like we need people need just get started I think it's we don't have a lack of very well qualified people in Europe in this phase
of I I there a lot of you know greatly qualified people but we still have much fewer entrepreneurs so a lot of the people then go and more koshy jobs and I think so encouraging more people right out of university but also right out of professional jobs that have been fermenting for many years and encouraging in them to actually start companies I think is super important because that will also just create generally more buzz and if you would I don't know quadruple the number of companies you were billed per year and in AI and Germany not all of them will work out statistically speaking it will have it will have a positive impact and then I think the second thing is we need to attract more later stage venture capital it's it's very easy these they are but there's a lot of investors out there who do like angel and seed investments in AI but also in other areas but especially when it comes to like more growth stage investments so series ABC when it's
about five ten twenty fifty million of investment and we have those investors in Europe but they tend to focus more on less deep tech related topics and so they're very few investors in Europe who really invest in those topics and I think we need we need to draw more capital in there obviously it's great if the government invests there more but also we need to create incentives for large pension funds like it what happens in the u.s. to invest in these models we need to draw more capital from outside of Europe too because otherwise the communist state is healing at some point and then there will be a competitor in the u.s. or China who gets like five or ten times as much funding and as these companies are anyways all globally operating and will lose the race and then I think also that should that should be also the paradigm to to think about regulations like we need to build the category leaders of tomorrow in Europe because only then we will also be
in able to actually put them after our value systems and also regulate them because right now what we're trying with Facebook Google and regulating them in Europe it's it's really hard when they're pulled in a different geography with like a different mindset but also when their entities are somewhere else and there anybody in the operating company so I think building more of these companies in Europe would definitely help also in graining more of our regulatory and mindset here what you said is interesting that it has nothing to do about AI but more about entrepreneurship and startups more funding more to Pinellas regulations so the same problem almost verticals and just with just interesting because both of you did not mention data and that this is a good but regarding time before we I would like to ask some more questions we have here from the audience yes and this is coming from it's a really good question and it's just C and you he can maybe or he can say what this means how can
a I helped solve great or the grand challenges we are facing post corona society climate change etc short answer please you know 30 seconds each well I think short answer would be it will probably help us to understand two problems much better because we can look at a lot of data much more than the human brain can process to understand what would be the right ways to tackle these problems in in the sense of what are the root causes for these problems and how to address them maybe maybe maybe for climate change there is a really good paper by I think yang dakuten and I don't know yoshua bengio is also on there but like basically all of the kind of leading AI figures around the world which brought this pretty long paper it's a few dozen pages around tackling climate change with AI and that paper has I think like 50 different things that cook you take away and also trying to rate okay doesn't really work yet what is it what is really
the impact and so I think that is a perfect starting round to look at this that was good Trotter question why doesn't male antics have a b2c product well that's a great question I think it's not it's not that we have you know deliberately decided against b2c space it's just that we have so far not found an interesting model we would want to invest in in the end be to see a space and maybe part of that is that building a successful b2c company is usually much less about technology than it is about doing good marketing and so and this is hiring and execution yes I'm timing an execution so I think again I I could well imagine ourselves going into into a b2c case if you find one but also it's really fair to say that such a case usually needs very different things than the you know the perks and expertise we have and can provide ok from my view it's by the way the b2c gains is gone because Amazon Google Facebook
they already own all the end-users and the attention and it's almost impossible to take away their market share to get their attention versus just building on top and then you don't have a billion-dollar company so I agree it's it's not easy but the whole b2b business I mean we speak about the whole industry they are not tackled there and here you're competing with Google and others from zero so the b2b market for AI is way larger moment and a quick question for you from Rasmus from robben he asked which cloud environment for AI development would you recommend for early-stage ventures yeah I think we've been working with all the large cloud providers so Amazon Google and Microsoft and also telecom I think they are they are all fine I think you need to I think we decided deliberately on against on-premise solution just because we figured yeah if you're just a few people it's fine to build by your own GPUs and everything on-site but once you scale and have a ton of people accessing
the data and the data needs to be safe and there's a lot of resources you need a cloud environment it's just more flexible so that's what we decided on day one and but in the end also most of our code runs on various of these environments you sometimes need to look for some industries if you if you if you using if your if your various industries which are very sensitive around their data and not for legal reasons were more from a perception perspective it helps to go with the European cloud provider or some solution which is maybe even else in Germany cuz like we've seen that sometimes ink and sales processes it makes it a bit easier but if you are you in a less critical area I mean any of these work and you can choose whatever is best from from a technology one of the very most difficult questions but I give you again 30 seconds max to answer because has some says some call it data mining some call it deep learning
some calls machine learning what do you really mean what AI is let's say host what is your definition of AI today this is always the hardest question yeah very big disco definition internally I mean it's just about creating a machine that makes intelligent decisions and so if you would show that to your new customers and they would be like hey this this machine is smart oh this is surprised by their decision-making then I think you'd probably call it AI but like on a more like technical level I mean we do a lot of deep learning on computer vision we do a lot of deep learning also for language understanding we would call those AI probably I mean the more technical term and also how we mostly call it internally is machine learning or deep machine learning and to the outside I think to customers it's just sometimes easier to talk about it as AI and not talk about it as deep learning so that's but also we we make it a bit simpler for ourselves
but internally we try to use the more precise for its terms for this as well allium what is the I well you could you could go back to the touring right it's probably it's probably AI once you don't realize it anymore so yeah I think we mostly use machine learning because that's very that much more clear definition on that yeah you train a machine to recognize patterns in a large set of data and then come up with decision making process that is based on this data and and therefore then can act autonomously without intervention and it's either supervised or unsupervised in a training but in the end it takes decisions and the broader the more complex those decisions get become the more it goes towards a I like real intelligence right because if it's if it's just simple decision-making in a very narrow field we wouldn't call it AI you would call it simple machine learning a to us and and also me AI is really more towards you know a more holistic form of
intelligence that can also make decisions in new spaces where it hasn't seen a lot of training data and thus makes decisions okay last question and I asked you after our conversation maybe to stay one two minutes in the chat if there are more questions you can also see the polls there maybe hang around so last questions there from the audience is from David Stremme he's asking out of your ma top of your mind what is currently the strongest NLT company out there well what is in in technical or or business terms what you think so it's Google its Facebook for me but it's maybe something like digital genius it is pure AI driven what comes in your mind I mean I I think it's it's a large companies surely like so like a Google like a facebook also also like a Microsoft I think one company which is very strong and in Germany is tipo so in the in the language formation space I mean they're also constantly beating and the Google Translate and are
doing a great job so I would say within in Germany I would say there yeah the number one and there also it's not just the company where this great a technology but also they have a business around you right and Tom of users so I think it's a great example and something we can be very proud of before we finish um Wade east email antics in ten years dream for a second be bold you don't be German yep so I think our vision is is is to build more than 50 companies in the next 10 years okay 50 companies okay or more 50 AI driven companies or what we call AI later yet that's much more operational yeah and they think of another perspective like in if in 10 years we look at what are kind of the dominant companies in the globally with AI that hopefully we have at least one or hopefully several of them so mm-hmm actually have impact so tragedy buoy the pharmacy of woolen burger said the next trillion dollar
company will be an air company and he is expecting that like hardware like software like AI it will be tripling each time will be huger and bigger and I hope that from these 50 maybe handful of German unicorns will come from the Milan takes venture studio I I for myself because I've invested in you guys for me as someone I'm living in Berlin some are living in Europe we need more of this therefore thank you for joining today thank you for sharing a lot you know your knowledge with the live audience with the play big audience with the podcast audience everyone who will listen to this you will stay a couple of minutes in the chat there want two more questions you find them and you can answer them directly if you want you and within this as they thank you for joining today thanks a lot Fabian was a pleasure and that's really funny everyone else and stay on line for my closing words and saying goodbye to everyone next week we do not
have a virtual chat because it was planned that we have the Rise of AI conference next week see I didn't anticipate this that we have to cancel it when we can't plant everything so next week is a break about the week after um we have a talk with Daniel Jeffries keep in mind it's in mute date it's a Friday 22nd I think 22nd of May it's a Friday and maybe someone can link the the link here and we will talk about the post corona world we will talk what in P Corona has on society and business and technology then inject versus a futurist he's a thinker he's an author he's amazingly on Twitter and on Facebook like huge crowd follower he's a nerd he's a hackers programmer and he we both love to discuss for hours what if how will the world be so this will be a very interesting dialogue how the world changes under Corona and else I already have the new speakers for the next four weeks for you you will be
the first one today to hear what we have so next one being as a guest after Dan will be Chris post who is the founder of Aragon the CEO he is adviser to Angela Merkel the German Chancellor for AI he is in many boards and committees he's involved because he has expertise is really good at transporting knowledge this will be very interesting job and we speak about optimism like we've got the question today how can a I solve the big problems and then week after and you can sign up already all the time for them now and we have a lobbing handsome and Robin Hanson he's a professor at the Oxford University for Humanity where are where we also have the Institute from Nick Bostrom who wrote the book super intelligence and Robin Lee is author of the book of emulated Minds when Roberts ruled the world so if you've seen the serious upload on Amazon for example or have seen a couple of Moran Black Mirror episodes he wrote a book about this how
will the world look like if we uphold conscience of conscience but just our skill our intelligence and we have robots or I call them BOTS who started working for us but they increase you're not like instead of building in EIU build a human just artificial other way around I like the idea so we will discuss about the implications if you have a whole virtual work first doing a lot of stuff and then the last one I like to announce is Tomas Hyneman who's member of the German parliament and he will publish just couple of days before or book how to research Germany it's called moisture torch lamps so it's a German play and he has written a very good book with 15 other members of the Parliament what governments can do to make this a better place like cosmos and earlier already mentioned less regulation less bureaucracy more funding they wrote a book about this I really have read the book it is the first time that I get like 100 theses action items but
politicians could do what administration could do to make this work better so politics recommending for the politics how we can use a I hope we can use blockchain to make this world a little bit more seamless comfortable and successful saan how we find a European way in this so you were the first one to learn to know about this you can sign up online for them as usual it was a pleasure to have you here today I'm looking forward to having you there again in two weeks stay healthy stay at home and well we talk again